传媒精选 No.1743 传媒精选 No.1743

传媒精选 No.1743

传媒精选 No.1743

t.me/zwcmjx

ch001.xhtml

英伟达正引领人工智能热潮。好事一件

这家芯片巨头的豪赌，正是资本主义本该运作的方式

插画：蒂姆·麦克唐纳

2026年9月3日

1993年，英伟达（NVIDIA）以拉丁语中代表“嫉妒”的词汇命名。不出所料，这家美国商业巨头如今引来了无数嫉妒的目光。全球对其图形处理器（GPU）——驱动人工智能运转的核心芯片——那永难满足的需求，已将英伟达推上了全球市值最高公司的宝座，估值达5.4万亿美元。明年，它还可能成为全球最赚钱的企业，创造高达3700亿美元的净利润。到2029年，其销售额甚至可能突破1万亿美元大关。

作为英伟达掌舵人的黄仁勋，无疑是这场AI热潮中最核心的魔术师。自2022年底ChatGPT横空出世以来，该公司的股价已飙升至当年的14倍。在标普500指数的市值中，前十大押注AI的上市公司占据了40%的权重。仅英伟达一家就独占8%；而且与苹果或特斯拉这类兼做智能手机或汽车的企业不同，英伟达的赌注几乎全盘压在AI之上。自2023年以来，美股每带来一美元的投资回报，就有约15美分出自这家公司——正是这样的财富效应，让消费者在利率攀升、关税重压以及对伊战争的阴云下，依然保持着旺盛的消费力。

然而，许多人也将黄仁勋视为另一种令人担忧的魔术师，担心他不过是个正在吹大危险泡沫的幻术师。通过总额高达1万亿美元的一系列交易，英伟达向数据中心运营商和AI实验室提供资金或信贷担保，以便他们能采购自家的GPU。有人因此戏称其为“AI中央银行”。批评人士至少指出，英伟达的这种金融运作手段充斥着“供应商融资”的意味，而正是这种把戏在2000至2001年的互联网泡沫狂热中推高了思科等网络设备制造商的营收，其随后的崩盘更直接引爆了一场经济衰退。

不过，若退一步审视，这些担忧大多站不住脚。如果英伟达在AI上的巨额豪赌最终兑现，它们将加速这项技术的普及与应用，从而大幅提振生产力和生活水平；而一旦满盘皆输，代价也将主要由英伟达自身的股东承担。这正是资本主义理应运转的逻辑。

不可否认，当下的AI热潮与昔日的互联网泡沫确实存在令人不安的相似之处：一项令人振奋的新技术、一场史诗级的大牛市，以及一群傲慢自负的科技大佬。英伟达从一家向电子游戏玩家和加密货币矿工兜售芯片的小厂，一跃成为全球经济的关键命脉，其蹿升之迅猛，以至于许多人甚至还没搞清楚它的名字到底该怎么念（是“en-vidia”，而不是“nuh-vidia”）。这不禁让人联想起思科当年的腾飞轨迹——在2000年，思科也曾短暂问鼎全球市值最高的宝座。正如思科当年路由器和交换机的销售是建立在网络流量呈指数级增长的前提之上，如今英伟达的GPU营收，同样押注于市场对AI Token（词元）永无止境的需求。

思科对终极需求的判断并没有错，但它错估了时间节点——互联网泡沫由此破灭。而在今天，难以捉摸的恰恰是AI普及应用的速度。如果把那些被Claude迷得神魂颠倒的软件工程师撇开不谈，整个社会的实际应用依然处于初级阶段。倘若应用需求未能迅速爆发，英伟达的客户可能会在芯片销售额断崖式下跌的同时要求兑现担保。鉴于没人能确切知晓GPU折旧贬值的速度究竟有多快，英伟达收回的任何二手芯片届时都可能沦为一堆废铁。

英伟达的金融魔法在一定程度上也是出于防御。其最新的GPU产品早已无法在市场上独霸天下。英伟达大约一半的收入依赖于美国的云计算“超大规模服务商”，主要是亚马逊、谷歌、Meta和微软，而这些巨头目前都在自主设计芯片。非英伟达生产的AI芯片市场份额已从2023年的26%攀升至38%。为了维持领先优势，英伟达过去曾将超过五分之一的销售额投入研发，如今这一比例已降至不足十分之一。

最后，英伟达庞大的财务承诺规模确实令人心惊。摩根士丹利估计，到2029年初，该公司的整体信用敞口（即其适度的借款加上对客户的兜底支持）将达到2000亿美元。假以时日，英伟达的影子债务可能会攀升至3000亿美元甚至更高。这是一笔天文数字：在当今的美国，只有前六大银行背负的债务能超过这一水平。

然而，眼下这场热潮与昔日互联网泡沫之间的差异，远比其表面的相似之处更为关键。英伟达的资产负债表异常强健。该公司手握990亿美元的充沛现金，且吸金能力仍在不断增强。在过去的三年中，其年销售额几乎每年翻番。毛利率已从不足60%大幅跃升至75%。黄仁勋那教主般的CEO声望，如今已足以与埃隆·马斯克分庭抗礼。但不同的是，马斯克旗下的企业要么现金流微薄（如特斯拉），要么在大把烧钱（如SpaceX），而英伟达仅在今年就将创造大约2000亿美元的现金收益。

这意味着，当年思科依赖发债支撑，而英伟达完全可以用真金白银为其GPU买家的交易提供兜底保障。即便其承诺的担保全额到期兑现，且现金流从明年起停滞不前，到2028年，其杠杆率在标普500指数所有非金融企业中仍将处于极低水平，仅有39家公司的负债率能优于它。英伟达的利润必须在现有高位的基础上暴跌60%，才有可能失去其投资级的信用评级。

此外，对AI的需求并非镜花水月，绝不能与Pets.com等那些毫无营收的互联网泡沫宠儿同日而语。领军AI实验室Anthropic的季度销售额已从第一季度的50亿美元飙升至第二季度的115亿美元。作为其主要竞对的OpenAI，业绩恐怕也不相上下。各大云计算超大规模巨头也都在斩获真实的AI收入。综合来看，AI目前每年或已为美国科技行业带来约1500亿美元的收入，而就在几年前这一数字还是一片空白。虽然这距离覆盖2.5万亿美元的AI资本支出仍有相当差距，但其增长势头极为迅猛。

在黄仁勋看来，AI革命面临的最大阻碍并非需求疲软，而是基础设施的极度匮乏。既然公开市场无法提供所需规模的巨额资本，英伟达便亲自下场提供融资支持。在评估风险与回报的平衡方面，英伟达拥有无可比拟的洞察力。尽管这场豪赌的体量巨大到足以牵动整体宏观经济，但这本质上是一场属于企业家的商业冒险。任何一家选择将现金用于再投资而非回购分红的企业，其实都在赌自己能够跑赢市场大盘。企业的存在，本就是为了进行这种高度集中的战略押注；至于那些希望分散风险的投资者，完全可以自己在投资组合中做好配置。

况且，黄仁勋绝非孤军奋战。在英伟达崛起之前作为全球最成功典范的云计算巨头们，也在做着一模一样的押注。华尔街的顶级金融巨鳄亦是如此。上个月，高盛、贝莱德、黑石等知名机构纷纷联手英伟达，共同启动了一项规模达5000亿美元的数据中心建设计划。黄仁勋的股东们同样未见惊慌抛售离场。即便这一众精英全盘皆错，赌在赌桌上的也是他们自己的真金白银；而一旦他们押注成功，人类迈向AI时代的大门便将更快被推开。■

ch002.xhtml

唐纳德·特朗普的委内瑞拉交易大胆却暗藏玄机

更糟糕的是，它为阻碍民主化进程提供了利益驱动

摄影：美联社

2026年9月2日

石油曾让委内瑞拉富甲一方。数十年来，丰厚的石油收入维系着一个稳定的民主政体，并赋予其拉丁美洲最高的生活水平。1998年，左翼民粹主义者乌戈·查韦斯打着将石油美元重新分配给穷人的竞选承诺上台执政，当时该国日均原油产量达340万桶。然而，在他腐败而无能的治理下，汩汩流淌的石油与滚滚而来的财富几乎彻底干涸。而在他钦点的接班人尼古拉斯·马杜罗治下，政府更是肆意滥印钞票，引爆了恶性通货膨胀，生生逼得四分之一的委内瑞拉国民背井离乡。

八个月前，当唐纳德·特朗普派遣特种部队生擒马杜罗并将其押入布鲁克林监狱时，委内瑞拉民众曾欢欣鼓舞。如今，特朗普却与这位前独裁者的副手德尔茜·罗德里格斯达成了一项协议，双方均宣称该协议将重振委内瑞拉的石油工业。特朗普将其盛赞为“世界历史上最伟大的石油交易”。他洋洋自得地宣称，这将赋予美国“能源霸权”（此前他主动发动的对伊战争打乱了全球能源供应）。这笔交易确实极具魄力，但也同样丑态百出。

委内瑞拉政权已将坐拥约650亿桶原油储量的17处油田特许经营权，授予了私营石油企业北美蓝能伙伴公司（NABEP）——该公司原本就是委内瑞拉境内的第二大运营商。作为提供支持的交换条件，美国通过五角大楼获得了NABEP公司35%的股权及其产出配额权益。在未来一个世纪内，这确保了美国对占委内瑞拉总储量五分之一的油田拥有优先购买权，且部分采购还能享受折价优惠（详见美洲版块）。

图表：《经济学人》

此举的盘算在于：NABEP将提高石油开采量，而山姆大叔的入局背书，能够让其他投资者重拾信心，将长期资本注入这个他们此前避之不及的国家。如今已被特朗普认可为委内瑞拉临时总统的罗德里格斯女士声称，该项目最终有望实现日产超过150万桶原油。若真能如愿，委内瑞拉的总产量将推高至250万桶，恢复到查韦斯上台前四分之三的水平。

然而，她既未给出清晰的时间表，也未透露如此大规模的扩产所需的天文数字资金究竟由谁来掏腰包。更何况，这笔交易已亮起刺眼的红色警报。首先，该协议是与一个靠操纵最近一届大选上台的非法政权谈判达成的。其次，特朗普选定的私营合作伙伴——NABEP掌门人亚历杭德罗·贝坦库尔——在委内瑞拉部分民众眼中声名狼藉，因其凭借道德上的“高度灵活性”靠依附该政权牟取了巨额财富。第三，特朗普帝国主义式的狂妄言论必将激怒委内瑞拉人：9月1日，他公然宣称要将该国的所有石油尽收囊中。

委内瑞拉任何未来经由民选产生的合法政府，都难免会对这项赋予美国对该国石油如此庞大控制权的协议提出质疑。这一严峻现实必将令投资者举足倾顾。尽管原本就在委内瑞拉开展业务的雪佛龙刚刚宣布了一项独立的70亿美元投资计划，但多数石油巨头依旧保持审慎观望。

更为恶劣的是，这项交易所带来的巨额现金流与税收利益，为当权者阻挠委内瑞拉向民主过渡创造了现实动因。在加拉加斯与海湖庄园均拥有深厚人脉的贝坦库尔，完全有充分的理由向特朗普进言，鼓吹过早举行自由选举可能会砸了这门石油暴利的买卖。而罗德里格斯女士无疑也会在旁推波助澜。

在背景简报会上，美国官员坚称委内瑞拉迈向民主的过渡进程仍将继续推进。政府内部的部分核心人物，如国务卿马可·卢比奥，也清楚地意识到一个民主的委内瑞拉既符合道义期许，更切合美国的战略利益。因此，希望并未彻底破灭。从理论上讲，美国大可利用其对NABEP的控制权，推动委内瑞拉石油工业实现透明竞争。而这反过来将使反对派能够在不破坏该协议的前提下展开竞选活动。然而，这一切的变数全系于特朗普一人。令人不安的是，他在公开场合已经开始将罗德里格斯女士称为委内瑞拉的“当选总统”。■

ch003.xhtml

约韦里·穆塞韦尼的衰老正将乌干达推向险途

这个国家面临着沦入更残暴专制或陷入混乱权力争斗的前景

摄影：盖蒂图片社

2026年9月3日

在乌干达，当一名持不同政见者从来都不是件容易的事。自40年前率领叛军夺取政权以来，总统约韦里·穆塞韦尼在一个饱经血雨腥风的国家恢复了秩序，实现了强劲的经济增长，同时也残酷无情地镇压了所有对其统治构成的挑战。在他最主要的两名政敌中，博比·怀恩远走流亡；另一位奇扎·贝西杰则身陷囹圄并因叛国罪受审，今年7月他更是在法庭上当场晕厥。今年1月，穆塞韦尼在一场充斥着暴徒行径与塞选票舞弊的虚假选举中再度胜出。

然而，随着现年81岁的总统日渐衰弱，乌干达人开始不再那么担心他接下来的统治动作，转而忧心他撒手人寰后局势将如何发展。随着潜在接班人竞相争夺权位，乌干达的现状印证了一个残酷现实：当年迈的独裁者开始大权旁落时，局势往往会滑向极度危险的深渊。

穆塞韦尼曾以超长时间的高强度工作著称，但近来他在办公室露面的时间已明显减少。他屡次临时取消公务行程，在公开场合亮相时也显得步履蹒跚。尽管他在去年的竞选活动中曾试图在红毯上慢跑以彰显年轻活力，但这套作秀手段根本糊弄不了任何人。

随着最高领袖逐渐日薄西山，其身边的亲信集团正迅速填补权力真空。穆塞韦尼的弟弟萨利姆·萨利赫被普遍认为在幕后操盘着大部分日常政务。其他王室亲眷的势力也日益膨胀。最令人担忧的莫过于总统之子兼陆军司令穆胡齐·凯内鲁加巴。他曾自诩为未来的国家元首，甚至宣称自己是耶稣的后裔。随着他四处滥用职权、耀武扬威，乌干达境内的政治打压正变得愈发残暴且难以捉摸。

凯内鲁加巴将军以其在社交媒体上荒诞不经（且往往迅速删除）的言论而臭名昭著，其内容从炫耀性欲、种族歧视谩骂，到威胁要杀害或阉割反对派，无所不包。他的父亲过去尚能约束其乖张行径。在2022年的一条推文中，这位将军曾狂妄叫嚣“用不了两周，我和我的军队就能攻陷内罗毕”，这曾导致他被短暂解除军职。（肯尼亚是一个友好且体量更大的邻国，乌干达的出海口高度依赖其港口，穆塞韦尼随后不得不为这起对其首都的无端挑衅公开道歉。）

然而，随着年事已高，总统似乎已无力管束自己的儿子。今年6月，这位将军公然派遣军队占领了乌干达最大的私营媒体机构（在经历数周谈判后该机构才得以恢复运营）。他下令绑架持不同政见者。在8月下旬，他甚至指使手下绑架了自己姐夫的密友，对受害者的惨状冷嘲热讽，并公然威胁要将姐夫一并逮捕。

一旦穆塞韦尼离场，乌干达将何去何从？就连政府内部人士也开始心怀叵测地谈及局势可能重回20世纪70年代伊迪·阿明血腥屠杀式军事独裁的黑暗岁月。尽管这种极端倒退的可能性不大，但未来的前景依然一片黯淡。这位将军可能会子承父业，在乌干达确立一个反复无常、睚眦必报且带有种族色彩的专制政权。他还可能进一步加剧邻国刚果民主共和国东部的冲突乱局——乌干达过去曾频繁插手该地区事务，而他本人与刚果的主要死敌卢旺达私交甚笃。

穆塞韦尼核心圈子内的其他权贵以及军队高层可能会试图联手阻击这位劣迹斑斑的继承人。一场惨烈的权力争夺甚至军队内部的武装分裂恐将随之爆发。部分年轻军官对他效忠誓死，另一些人则对他的行径深恶痛绝；而当年参与内战的退伍军人虽对老总统忠心耿耿，对这位行事鲁莽的少主却充满警惕。

老人统治的国度

理论上，一位独裁者的离去或许能为自由选举撕开一道口子，怀恩或贝西杰甚至可能借此当选。但在现实中，这种可能性微乎其微。穆塞韦尼当年是在非洲民主复兴的浪潮中夺取政权的，彼时众多非洲国家纷纷告别军事独裁，转向多党民主制。在那段时期，西方援助国偶尔还会动用自身影响力来捍卫人权。

反观当下，军事政变肆无忌惮地频频上演，众多通过民主选举上台的领导人，反而处心积虑地寻求各种反民主手段以求长期霸占权力。乌干达深陷的泥潭应当给其同行敲响警钟：缺乏真正的民主选举来完成权力顶层的更迭换代，随着独裁者日益老去，动荡与灾祸必将接踵而至。■

ch004.xhtml

在许多国家，阻碍繁荣的最大樊笼筑在家庭之内

一个少些性别偏见的世界，将会富足得多

摄影：盖蒂图片社

2026年9月3日

放眼全球，女性长期以来都在为争取职场上的平等待遇而抗争。然而在某些国家，这场斗争却面临着更为严峻而基础的阻碍：她们首先要争取的，仅仅是迈入职场的资格。在一项实验中，研究人员为埃及的已婚母亲们在其住所附近提供了免费或廉价的托育服务，以便她们能够外出寻找工作。然而，最终接受这一提议的女性仅占11%。横亘在多数人面前的绊脚石，是那些固执地认为女人外出工作有伤风化的丈夫们。尽管高达90%的埃及女性并不认同这种观点，但沉重的传统桎梏却难以打破——更何况还要面对思想保守的丈夫那实打实的蛮横干预。

在中东、北非和南亚的大片地区，许多男性依然认定家族的声誉维系于将女性幽闭在家庭之中。他们唯恐妻子只要踏出自家后院的菜地半步去务工，便可能接触到毫无血缘关系的陌生男性，从而让整个家族蒙羞。这种陈腐落后的观念甚至往往借助暴力威胁来强制维系。在印度，针对女性的家庭暴力早已司空见惯，以至于近40%的女性自身竟然也认为丈夫在某些情况下殴打妻子是合情合理的。在马里，持有这一观点的比例更是高达近70%。

此类偏见正在让整个国家陷入贫困。在南亚，仅有三分之一的女性参与劳动力市场；而在中东和北非，这一比例更是低至五分之一。自1990年以来，在思想开明的法国，女性所获得的劳动收入份额已从35%攀升至44%；但在父权制根深蒂固的巴基斯坦，该比例仅从1.5%微幅提升至依然令人堪忧的9%。据世界银行测算，若能彻底消除女性就业面临的制度与文化障碍，许多国家的人均收入将提升五分之一——这一经济红利甚至超过了避免一场典型内战所挽回的损失。

仅凭这一点，就足以成为全社会向幽闭女性现象宣战的充分理由。但更深层的动因还在于：一旦女性拥有经济收入，家庭内部的权力天平便会发生根本性倾斜。《经济学人》在分析了涵盖27个国家的数据后发现，在对各国的性别偏见程度及家庭富裕水平进行控制调整后，拥有有偿工作的女性在各项家庭事务中都享有更大的话语权——从支配家庭收支预算，到探访友人是否需要事先征得丈夫同意，无不如此。简而言之，有偿劳动让家庭关系变得更加平等。

因此，在所有将女性紧锁于门内的国家，推开这扇“玻璃大门”应当被列为当务之急。只可惜，这扇门上挂着数道错综复杂的重锁。需要插入的第一把钥匙是法治建设。据世界银行统计，全球有5.4亿女性所生活的国家中，法律面前的形式平等不仅存在瑕疵，甚至称得上是“遥不可及”。然而，即便是最根深蒂固的性别歧视法律也并非不可废除，沙特阿拉伯的蜕变便是明证。在2011年启动一系列改革之前，那里的女性几乎被禁止涉足绝大多数工作岗位，甚至被剥夺了驾驶汽车的权利。而如今，她们不仅能够自由工作、考取驾照，甚至可以自主选择是否佩戴头巾。自2010年以来，沙特女性的劳动参与率几乎翻了一番，从18%跃升至34%。对于一个曾被外界普遍认为在陈规旧俗中积重难返的王国而言，尽管前路依然漫长，但这无疑已是惊人的跃进。

第二把钥匙在于重塑社会规范。人们常认为社会习俗积习难改，但历史印证了其转变往往可以在旦夕之间。直至1975年，英国的银行仍有权在缺乏丈夫书面同意的情况下拒绝为女性发放房屋按揭贷款；直至1997年，德国法律才正式将婚内强奸定为刑事犯罪。自那时起，英德两国的社会文化经历了深刻洗礼，如今此类歧视性规定已到了令人匪夷所思的地步；事实上，当今年轻一代的英国人或德国人中，几乎无人知晓这些荒谬的历史曾真切存在过。

社会观念的转变既可以依赖自上而下的政策推动（如沙特的实践），也可以通过社会群体的示范效应逐步发酵。面对道德伦理问题，极少有人会从第一性原理出发进行严密推演。在大多数情况下，人们往往是在目睹身边大量相识且受人尊敬的人身体力行之后，才认同某种行为是合情合理的。因此，敢为人先的拓荒者至关重要：在巴基斯坦保守的偏远村落里，第一位顶住风言风语毅然前往邻近城镇务工的女性，便为后来者铺平了前行的道路。学校、非政府组织以及政策制定者同样可以在这一进程中推波助澜。在印度开展的一项实验表明，若能让男性亲眼看到他们相熟的女性在当地一家整洁而安全的工厂中稳定工作，他们便更有可能默许自己的妻子外出务工。任何旨在改善工作安全环境的举措——例如印度为夜间通勤女性开通的专属女性巴士——都能逐步瓦解那种“为了人身安全必须将女性幽禁在家中”的认知谬误。

数字科技的发展同样能够赋能这一变革。借助手中方寸大小的手机屏幕，身处极度压抑国度中的女性得以窥见——并进而勾勒——一种全然不同的生活蓝图。这并不意味着她们必须全盘照搬高度自由化的西方社会。在更多时候，她们只需借鉴那些性别观念稍显平等的邻国经验即可，例如当一位伊拉克家庭主妇在土耳其肥皂剧中看到女性身居高位、受人尊敬时，她便会不由自主地开始反思并审视自己的人生处境。

最后一把不可或缺的钥匙则是金钱。正如再保守顽固的男性也不得不承认的那样，财富具有无可替代的现实效用。一位经济学家曾提出过著名的“体面与收入的权衡取舍”假说：在某个临界点上，外出工作的妻子所能带回的丰厚收入会变得极具诱惑力，以至于做丈夫的会暗自权衡——放手让妻子外出务工，难道真会对所谓的家族体面造成不可挽回的毁灭性打击吗？就这样，尽管拜金主义在各类宗教典籍中常遭唾弃与鞭挞，但在现实世界里，对金钱的渴望却实实在在地成为了推动社会破旧立新的前行动力。■

ch005.xhtml

贝叶挂毯给人类逆境求生带来的启示

一幅拥有950年历史的亚麻刺绣，为现代人提供了弥足珍贵的历史镜鉴

摄影：诺曼底遗产制造工坊

2026年9月3日

1944年巴黎解放前夕，希特勒麾下的党卫军头目海因里希·希姆莱发出了一份绝密的加急电报。他在电文中急切地告诫巴黎的纳粹指挥官：“切记将贝叶挂毯转移至安全之地。”而那位焦头烂额的指挥官当时显然有更多性命攸关的燃眉之急需要应对。纳粹党人对这件描绘诺曼征服英格兰史诗的艺术杰作近乎痴狂，将其奉为雅利安人无上荣耀的至高图腾。希姆莱原本筹划将其陈列于维威尔斯堡——一个被他寄予厚望、企图打造成“新世界中心”的党卫军秘密大本营。然而，英国密码破译专家成功截获了这道指令，并迅速转交给了法国抵抗运动组织，后者最终神不知鬼不觉地将这幅传世挂毯秘密藏匿于卢浮宫的地下密室之中。

图拉真留下了巍峨的纪功柱，拿破仑铸就了雄伟的凯旋门。相比之下，诺曼底公爵威廉的这幅刺绣挂毯虽不及前者气势磅礴，却在历史的长河中显得更为令人难忘。其连环漫画般的叙事手法诙谐明快，极具现代“模因”特质，曾被后人广泛借用来讽刺从脱欧谈判到安迪·伯纳姆“北方征服”等形形色色的政治事件（在《新政客》杂志的封面上，那位被伯纳姆取代的前英国首相基尔·斯塔默爵士，便被刻画成双眼插箭的形象，活脱脱就是当年那位倒霉透顶的哈罗德一世国王）。如今，这幅挂毯自近千年前可能由坎特伯雷战败的盎格鲁-撒克逊工匠刺绣完成以来，首次重返英格兰故土。它将于9月10日起在大英博物馆公开展出。对于身处现代社会的观者而言，这幅历经浩劫的巨作至少揭示了三则生存启示。

首要的生存之道在于：若想在绝境中求得一线生机，有时最好的策略便是切忌锋芒毕露、引人注目。在长达数百年的岁月里，这幅挂毯据信一直被卷曲封存在一口沉重的木箱中，唯有在每年一度的特殊宗教节日里，才会被请出陈列于贝叶大教堂的中殿之中。在其存世的大约三分之二的时间里，它在外界几乎毫无知名度可言。直到18世纪初叶，更广泛的公众群体才逐渐注意到它的存在。正是这种长期不事张扬的隐秘状态，为它的周全保全筑牢了安全屏障。

质朴无华的材质同样功不可没：它仅仅是由普通的亚麻布与羊毛线编织而成。许多中世纪的华贵壁毯往往极尽奢华，通篇织入金线银丝，或镶嵌以无数璀璨夺目的珍珠宝石，以便在摇曳的烛光下折射出耀眼的光芒。如果后人未曾听闻或目睹过其中的绝大多数，正是因为它们早已湮灭在历史的尘埃中——昂贵奢华的装潢使它们极易沦为宵小窃取或拆解变卖的目标。即便是置身现代社会，炫耀珠光宝气依然难免招致无妄之灾。就在去年，法国部分存世的王冠珠宝在卢浮宫内被盗；上周，一条珍贵的皇家钻石项链又在维也纳的一家博物馆中失窃。然而，千百年来却从未有一只窃贼的黑手伸向贝叶挂毯。（当然，客观而言，要想神不知鬼不觉地盗走一件长度堪比波音747客机的庞然大物，其难度确实非同小可。）

第二则启示在于：在逆境之中，身旁若能有一群尽职尽责的能人志士为你挺身而出，往往能够逢凶化吉。法国大革命期间，当激进的革命军打起这幅挂毯的主意、企图将其充当运送军火弹药的马车防雨布时，全凭一位法国律师苦苦哀求，并自掏腰包搜集替代材料进行行贿周旋，才最终虎口夺食将其挽救下来。普法战争爆发之际，守护人员将其牢固密封并深藏于特制的锌制密封圆筒内，使其在战火硝烟中毫发无损。二战期间冒着生命危险从纳粹魔爪下保全这幅挂毯的法国抵抗运动战士们，尽管未曾亲历诺曼底海滩的浴血拼杀，却在无形之中为捍卫诺曼底的文化遗产打赢了一场惊心动魄的保卫战。

第三则深层启示则根植于挂毯本身所传达的叙事母题之中：那是一种属于胜利者的宽仁胸襟。从本质上讲，这幅挂毯无疑是一部战争政治宣传画卷，极有可能是由厄德主教亲自出资委托制作，旨在为其同父异母的兄长威廉在1066年黑斯廷斯战役中建立的赫赫战功歌功颂德。然而令人称奇的是，这幅作品对战败的英格兰人却展现出了异乎寻常的体恤与尊重——或许是因为战役结束后英格兰各地叛乱烽烟四起，在伤痕累累的败军伤口上肆意撒盐绝非长治久安之道。挂毯并未刻意掩盖或粉饰战争的血腥与残酷；但也绝未流露出哪怕一丝小人得志的幸灾乐祸，从未在失败者的坟头狂欢起舞，更未将英格兰人轻蔑地贬损为卑贱的“失败者”。

那些狂妄追逐虚名、热衷于将个人纪念碑堆砌得金碧辉煌、对异邦民族充满傲慢鄙夷、且无休无止地吹嘘那些从未真正赢取过的虚幻胜利的当代统治者们，应当在历史的镜子前深自反省。光阴流转近千载，威廉公爵在历史的丰碑上依然被尊称为一代“征服者”。而当今那些不可一世的强权巨擘们，在曲终人散、归于尘土之时，又能给后世留下怎样的历史盖棺定论？■

ch006.xhtml

对数据中心产生的道德恐慌实属荒唐

一场由虚假信息与认知误解共同助推的群体性抵制

摄影：盖蒂图片社

2026年9月3日

资本供给的短缺绝非制约美国AI巨头大举投资的唯一瓶颈。他们如今还不得不面对棘手的政治阻力。伴随着愈演愈烈的怒火，越来越多的美国选民正群起反对兴建数据中心——而这些基础设施正是训练和运行庞大AI模型所必需的算力基石。这场抵制狂潮跨越了党派鸿沟，从民族主义右翼蔓延到了激进环保主义左翼。其反弹声浪之大，已然使基础设施建设沦为今年中期选举中最具决定性的核心政治议题之一。政客们竞相角逐，争先恐后地抛出针对数据中心建设的种种限制、缓建甚至全面禁令。然而，选民与政客们都在犯下一个灾难性的战略错误。

流传甚广的一个典型谣言，是指责数据中心是吞噬水资源的无底洞，其耗水量被夸张到堪比印度神话中一口饮尽大海的投山仙人——这番耸人听闻的言论在很大程度上得益于某本畅销书中错漏百出的粗糙估算。而在现实中，一座中等规模的数据中心的耗水量大致仅相当于两个高尔夫球场；如果采用如今已被广泛普及的循环水冷却技术，其实际水耗还要低得多。更何况，根本没有任何合理依据将矛头单单对准AI产业：在现代生活的方方面面，从在Netflix上流媒体看剧到随手开启洗碗机，其耗水规模在那些苦修派环保主义者的放大镜下，都可以被轻易扣上穷奢极欲、肆意浪费的帽子。谷歌经过严密测算指出，处理一条Gemini文本指令的耗水中位数大约仅为五滴水；这意味着哪怕连续进行数万次AI交互，其累计消耗的水资源也抵不上站在淋浴喷头下冲凉一分钟的水量。

相比之下，公众对电网负荷激增的焦虑似乎更具数据支撑：数据中心的运转确实需要耗费极其惊人的电能。然而，迄今为止几乎没有任何确凿证据表明数据中心的引入直接推高了民用电价。恰恰相反，伴随着电力总需求的急剧扩容，固定基础设施的投资建设成本被分摊到了更多被售出的千瓦时电量之中。就在最近几周，多家大型公用事业电力公司纷纷证实，数据中心的集中接入实际上促成了居民电费账单的边际下降。不仅如此，众多数据中心还在积极出资，直接反哺并资助新型清洁能源发电项目的落地。

不可否认，AI的全面普及在未来确实有可能对电费构成上行压力。但面对不同行业对有限资源的争夺——无论是水、电力还是其他生产要素——最理性的应对之道在于根据资源的稀缺程度进行市场化定价，从而在长期维度上倒逼更充沛的资源供给并带动整体生活水平的持续提升；而不是依据道德卫道士们武断的主观臆断去强行搞要素配给。在诸如噪音污染等市场定价机制难以精准覆盖的领域，合理的规划管控固然不可或缺；但除了必须依托主干电网布局之外，原本就没有任何理由非要把数据中心生搬硬套在居民住宅区身旁。退一步讲，即便是关于噪音的焦虑也往往被严重夸大了：在网络上疯传的那些所谓AI数据中心产生刺耳轰鸣的视频，经查证绝大多数都是粗制滥造的比特币矿场，根本不能代表那些由超大规模云厂商主导建设的现代化尖端设施。

与那些被凭空捏造出的所谓沉重代价相比，数据中心带来的实际红利却是实实在在的。在国家战略层面，数据中心承载并驱动着一项全球需求正呈井喷式爆发的前沿核心技术，未来的宏观经济增长动能乃至国家安全防御壁垒都将深系于其成败。而在更为微观的现实层面，数据中心向地方财政缴纳着巨额税款，直接充实了用于资助公立学校、社区图书馆和警察局的财政资金。在弗吉尼亚州的劳登县，其境内250多家数据中心内部署的巨额计算设备所产生的财产税，预计将在明年贡献全县财政总收入的40%。这笔飞来的财税横财，使得该县得以将普通居民的房产税税率直接下调了约30%。

主张放缓AI前沿技术演进的研究逻辑并非全无道理，但其立论的核心应当且必须建立在该技术本身可能被恶意滥用或彻底失控给全人类带来的存在性危机之上，而非纠缠于数据中心在局部地区所造成的鸡毛蒜皮式的所谓危害。美国若单方面自断经脉、人为制造阻碍限制本土AI基础设施的建设，无异于将战略竞争优势拱手让给中国，而后者的威权领导层完全可以随心所欲地在任何选定地点迅速铺开数据中心集群。多年以来，美国的民粹主义政客们始终在哀叹实体经济的空心化让中国抢占了经济先机，并导致山姆大叔在战时面临关键工业产能短缺的困境。如今，他们对数据中心建设的横加阻挠，恰恰是在亲手重蹈覆辙，制造出他们昔日最为痛心疾首的同等战略灾难。■

ch007.xhtml

欧洲无法阻止俄罗斯走向自我毁灭

拉德克·西科尔斯基撰文指出：西方对弗拉基米尔·普京彻底走向疯狂的恐惧，已经让自身瘫痪了太久

插画：丹·威廉姆斯

2026年8月27日

“世界变得更健康了，但并未变得更安全。”1953年约瑟夫·斯大林逝世后，《经济学人》曾留下这样一句沉痛的断言。当时，心怀这般忧虑的绝非仅此一家媒体。在人们眼中，某种人尽皆知的俄式极权主义，似乎无论如何也比不可预知的未来更具可控性。弹指一挥七十年过去，这种弥漫在西方世界的深层心态却几乎未曾改变。无论克里姆林宫做出何等出格的举动，外界对于“一旦对其施加坚决反抗，对方恐将采取更加不计后果的疯狂行径”的莫名恐惧，始终被频频拿来当作向俄罗斯霸道诉求步步妥协退让的遮羞布。

且看本刊最近刊发的一篇对俄罗斯寡头安德烈·梅尔尼琴科的深度访谈。他坦承弗拉基米尔·普京确实犯下过一系列战略失误，但随即力辩称：接下来的局势演变只可能更为凶险万分；俄罗斯倘若遭受过度极限施压，极有可能“彻底沦入万劫不复的动荡泥潭”，乃至“铤而走险动用战术核武器”。其潜台词再直白不过：要么由着俄罗斯在国际舞台上肆意妄为，要么大家一同承担玉石俱焚的灾难性后果。这完全是一条饮鸩止渴的死胡同。

在苏联轰然解体后的漫长岁月里，西方世界曾竭尽全力避免让俄罗斯遭受大国陨落的屈辱，处心积虑地试图将其塑造成国际秩序中负责任的利益攸关方。仅在1997年这一年内，莫斯科便受邀列席G7峰会，与北约共同签署了具有历史意义的《北约与俄罗斯相互关系、合作与安全基本文件》，并与欧盟缔结了《伙伴关系与合作协定》。在后苏联时代，历届美国总统政府都曾试图通过签署各类军控条约、深情凝视俄国领导人的双眸、或是煞有介事地按下所谓的外交“重启键”，试图将俄罗斯拉回文明世界的怀抱。曾沦为苏联卫星国的诸多东欧国家亦复如是，这其中便包括波兰。

2009年，普京曾应邀亲赴波兰，出席纪念第二次世界大战爆发70周年的国家公祭仪式。我的祖国当年曾为毗邻波兰边境的俄罗斯加里宁格勒飞地居民量身推出了免签过境政策。俄罗斯东正教会与波兰罗马天主教会的最高领袖，更曾联名发表历史性倡议，呼吁两国人民“相互宽恕彼此在历史上所遭受的委屈、不公以及一切罪恶”。而我自己，仅仅因为曾公开表示“只要俄罗斯能够完全符合各项准入标准，我完全可以设想一个真正走向民主化的俄罗斯加入北约组织”，至今仍在波兰国内遭到在野反对派毫不留情的政治围剿与人格攻讦。

我们曾经倾尽所有去尝试接纳。然而，我们所收获的却是一次又一次冰冷无情的拒绝。

时至今日，俄罗斯再度故技重施，企图通过挥舞升级冲突的大棒来强行达成其战略图谋。它在近期公然威胁英国，声称若继续向乌克兰提供尖端军事援助必将面临“严重后果”。今年8月，美国中央情报局局长约翰·拉特克利夫秘密出访莫斯科，据传其核心使命便是当面向克里姆林宫发出严厉警告，勒令其切勿对任何北约成员国轻举妄动。

然而，尽管这位俄罗斯强人除了一味勒索乌克兰彻底屈膝投降之外，根本未曾流露出任何寻求公平和平解决方案的诚意，国际社会上依然有太多政客对这一铁一般的事实视若无睹。一些人主张，在日益加剧的经济制裁与战场损耗的高压之下，俄罗斯已经变得虚弱不堪，西方必须赶在莫斯科政权轰然垮台之前、或者更糟糕的——赶在俄联邦彻底四分五裂并催生出一批比普京更为极端凶险的继任狂徒之前，尽快开启停火谈判。另一些人则固执地坚称，决不能将俄罗斯逼入绝境，否则它可能会诉诸全方位的战事升级，包括突破核战争的红线。这两种貌似深谋远虑的说辞，实则全盘误判了形势。

对于持续施加战略高压将导致俄罗斯国家陷入解体解构的担忧，完全是被严重夸大了。俄罗斯高度集权的政治架构，早已使其各个联邦主体在经济命脉、行政运作和财政转移支付上深度依附于莫斯科。数以千万计的边疆国民在生计上高度依赖联邦中央财政的转移支付救济，即便是资源富庶的边远地区，也同样受制于由中央牢牢垄断的全国交通网络与能源输送管道。今日的俄罗斯绝非昔日的苏联帝国。除了局势复杂的北高加索局部地区之外，俄罗斯族人口在绝大多数联邦领土上占据着绝对压倒性的主体地位，加之当年伴随苏联解体而来的经济大崩溃与全面政治动荡所造成的惨痛社会创伤，至今仍在强力抑制着任何实质性分离主义思潮的抬头。更为重要的是，这个国家在长达数百年的历史积淀中始终延续着深度中央集权的统治基因。即便高层爆发严重的统治合法性危机，抑或在对外战事中遭遇惨重挫败，最终最可能引发的也是统治精英阶层内部围绕夺取中央中枢权力的惨烈宫廷政变，而绝非庞大版图的分崩离析。

至于那些或由莫斯科官方渠道直接咆哮而出、或通过所谓“忧国忧民”的寡头喉舌隐晦放出的升级冲突威胁，绝非俄罗斯对新兴地缘危机的应激反应，而不过是其沿用已久、驾轻就熟的经典讹诈剧本中的既定篇章。在今天，正如在漫长的历史长河中一样，俄罗斯始终习惯于动用暴力手段来反复测试西方的意志底线。冷战时期的苏联就曾反复利用有限烈度的试探性进攻，来侦测对手的战术虚实与心理防线，随后再相机发动全面攻势。今天的克里姆林宫在战略战术上依然在原汁原味地套用这套旧逻辑。在当前向西方民主世界发动的隐秘战争中，它大肆动用自杀式无人机骚扰、毁灭性网络攻击、跨国破坏活动、大面积民航GPS信号干扰以及铺天盖地的认知虚假信息战。对于西方阵营而言，核心问题早已不再是俄罗斯是否正在试探北约的战略底线，而是北约自身究竟是否清醒地意识到这种无休止的挑衅与试探，以及究竟准备对外发出何种坚决有力的反制信号，以遏制这种“灰色地带”的蚕食行动彻底蜕变为全方位的大规模正面对抗。

克里姆林宫的全部冒险豪赌，完全建立在一个极其简单的心理预期之上：它并不自欺欺人地认为俄罗斯自身实力有多么强大，而是笃定对立的民主阵营在面对危机时必定瞻前顾后，绝不敢对其致命软肋发起实质性反制。因此，欧洲必须坚决抵制任何开启过早外交妥协谈判的危险诱惑。那些大声疾呼重开政治对话之门的人，或许一厢情愿地寄希望于外交斡旋能够最终缓和地缘对抗的紧绷神经。然而从长远来看，事实恰恰完全相反。在俄罗斯的整体战略态势尚未被彻底削弱打痛之前草率开启妥协对话，只会进一步巩固克里姆林宫在其国内的极权统治合法性，助长其继续肆无忌惮地推进混合战争冒险，严重动摇并瓦解欧洲对乌克兰至关重要的战略援助决心，并最终全方位加剧西方民主阵营内部的离心倾向与战略分歧。

一个令人倍感不适却又无可回避的严酷真相是：欧洲必须在当下承担起更大的短期战略风险，唯有如此，方能彻底规避未来在泥潭深处面临的更为毁灭性的全面绝境。我们的战略目标绝不能仅仅停留在消极防范冲突升级的狭隘防线之上。我们的真正目标，必须是让普京彻底认清现实：任何升级讹诈手段都将注定一无所获，并且我们完全准备付出远超金钱层面的巨大代价，去坚定捍卫我们自身的生存与尊严。欧洲各国领导人肩负的首要天职，是全力保护自己的国民免遭俄罗斯帝国主义野心的真切残害，而不是倒戈去保护俄罗斯免遭其自身宿命的吞噬。

俄罗斯早已在意识形态上将我们视作不共戴天的死敌，并不遗余力地企图彻底埋葬我们的民主文明。我们绝不可在恐慌中急于对其绥靖退让。当普京真正准备好拥抱和平的那一天，他自然会找到传递这一意图的方式与途径。■

拉德克·西科尔斯基系波兰外交部长兼副总理。

ch008.xhtml

全球在气候变化问题上的自救时间可能远比想象的更为紧迫

蒂姆·帕尔默撰文警告：如果气候变化正在颠覆决定反气旋生成机理与路径的大气环流模式，全球气候危机或将以惊人的速度恶化

插画：丹·威廉姆斯

2026年8月30日

对于我所定居的英格兰南部而言，今年夏天的极端恶劣天气用“残酷”二字来形容绝不为过：罕见的极端高温热浪席卷大地，打破历史极值的干旱少雨纪录更是接踵而至。铁轨在烈日暴晒下发生严重扭曲变形，致使大面积交通网络陷于瘫痪；农作物产量出现断崖式暴跌，农场主们已然纷纷发出严厉警告，明年的粮食供应恐将面临严重短缺。

一些气象专家轻描淡写地告诉公众，这一切不过是长期以来气候变化预测模型中关于“夏季更热、更干燥”趋势的常规体现。然而，此类宽泛的概念性预测根本无法从本质上解释我们今年夏天所亲历的极端气候异象；而如果我们想要为明年夏天以及随之而来的数个漫长苦夏做好充足准备，深入——且刻不容缓地——厘清其中的真正症结，便具有压倒一切的紧迫性。

那种公众耳熟能详的“暖湿之冬、干热之夏”的经典气象预测，在专业领域内通常被称为气候变化的“热力学”特征：即由于温室气体排放不断加剧，全球地表平均气温必然随之攀升，而愈发温暖的大气层中必然能够容纳更为充沛的水蒸气，这是最基础的物理规律。在大气湿度大幅增加的背景下，每逢冬季为不列颠群岛输送丰沛雨水的大规模常规低压系统，在气候变化的推波助澜下必然会带来更为猛烈的极端强降水；相反，到了炎炎盛夏，伴随着太阳直射点的持续北移，赤道与两极之间的温差梯度被大幅压缩，使得上述降水低压系统受到严重抑制，温度更高的空气便会加速蒸发土壤水分，从而导致降雨量进一步急剧萎缩。我们对于这一热力学信号机制的认知已然十分成熟透彻——这在很大程度上得益于其背后的物理学机理极其清晰简洁。

然而，上述这套经典理论与今年夏天席卷欧洲的罕见灾难之间，并没有任何直接的因果关联。导致这场大旱灾的真正罪魁祸首，是一个自晚春时节便死死盘踞在西欧大部地区上空久久不肯散去的持久性超强高压系统——即反气旋。正是这个强大的异常反气旋，如同一面坚不可摧的无形力场盾，将所有本应途经欧洲的降水锋面云系强行阻挡在外。要想彻底探明今年夏天何以如此酷烈焦灼，首要任务便是揭开这个超长待机的异常反气旋背后的深层动力学成因。

粗略归纳起来，科学界目前主要存在两种推论方向。第一种可能性在于，这不过是地球气候系统内部混沌无序的自然年代际波动的一部分，与人为诱发的气候变化并无直接牵连。倘若这一推论属实，那么在即将到来的2027年夏天，再次遭遇如此顽固反气旋肆虐的概率将微乎其微。换言之，明年夏天我们有望迎来相对“正常”的大气环流模式。当然，即便是常规环流模式回归，气候变化的热力学印记依然无法抹去：与20世纪的夏季均值相比，2027年夏天预计仍将更为温暖干燥。但无论如何，公众至少不必再去承受今年这般毁灭性的持续大旱与连绵酷暑的折磨。

然而，第二种可能性则要令人背脊发凉得多：即这个赖着不走的异常反气旋本身，恰恰是气候变化推波助澜下的直接产物。倘若这一假说被证实，便意味着气候变化不仅在微观层面上深刻重塑着大气的热力学热力状态，更已然在宏观维度上全面颠覆了大气的“动力学”平衡。这是一个至关重要的分水岭。在区域气候研究尺度上，如果仅仅将气候系统片面地比喻为一台巨大的原始蒸汽机——仿佛只要人类不断往炉膛里铲入化石燃料，机器内部便只会单调地不断升温加湿——这种认知显然过于幼稚；我们必须精确追踪并严密预测构成整个庞大复杂气候系统的各个关键机械传动部件——特别是大气圈与水圈——其内部各个齿轮与轴承之间的受力状态与运动规律。

在第二种灾难性的情景预设中，气候变化正在全方位重塑大气环流的基本格局，甚至包括直接左右反气旋生成与路径迁移走向的平流层极锋急流。现有的大量气象观测实证数据表明，在盛夏时节，此类持久性反气旋的爆发频率确实在显著攀升，并且呈现出与全球平均气温升高高度同步的正相关性。然而，想要通过数字模拟模型完整重现并印证这些观测数据，却是一项极其艰巨的科学挑战：与相对线性的热力学机制截然不同，气候变化的动力学机理呈现出极度复杂、高阶非线性的动力学特征，唯有依托当今世界最前沿的尖端超算气候模型，才有可能对其展开定量的严密推演。

若将这一理论推导至极端情形，我们可能正飞速逼近甚至跨越一个恐怖的临界点：在这一拐点之后，气候变化的动力学反馈效应将变得无可遏制，持久性极端反气旋或将顺理成章地沦为今后每一个盛夏的常态化配置。如果这正是全人类眼下所面临的残酷现实，我们将彻底丧失在未来一二十年内从容适应气候“新常态”的宝贵时间窗口。我们必须在瞬息之间完成全方位的应激适应。

部分气候建模专家或许会指责我在此危言耸听。毕竟，即便是当今世界上最庞大完备的全球气候超级模型，也并未推导出持久性反气旋将必然成为未来每个夏季常客的预测结论。然而我们同样心知肚明的是，现有的这些尖端模型在模拟今年所观测到的这种超强持久性反气旋时，表现得极为挣扎吃力——这在很大程度上是因为它们在空间网格与时间步长分辨率上严重不足，根本无法精细表征并捕捉大气层中不同物理尺度之间错综复杂的湍流级联相互作用。而众所周知，正是这些尺度间的复杂相互作用，才是维系反气旋系统长期盘踞不散的核心动力来源——这恰如木星表面那场肆虐数百年的大红斑风暴一般。现有数值模型未能成功重现这一潜在致命临界点，绝非我们高枕无忧、盲目乐观的借口，而恰恰是倒逼我们必须不惜代价大幅升级气候模型的震耳警钟。

更为严峻的变数还在后头：新一轮厄尔尼诺现象已然在太平洋深处积蓄力量。对于英国本土而言，热带太平洋海域的这一周期性异常偏暖现象，其最直接的冲击往往集中在隆冬时节——即将到来的这个冬季起初可能会比往年更为阴雨连绵，但在后半程却极有可能转入异常反常的干燥少雨模式。然而，欧洲的天气系统在应对不同强度的厄尔尼诺事件时，其波动幅度极为剧烈莫测。倘若这一轮厄尔尼诺在今冬为英国带来的实际降雨量显著低于历史均值，各大主力水库将根本无法在明年入夏之前完成蓄水补给，随之而来的次生灾害后果将不堪设想。

面对死局，人类究竟该何去何从？多年以来我始终在各类国际场合大声疾呼：全球各主要大国的国家级气象科研机构必须立即打破藩篱，全方位整合顶尖科研人才与超算算力资源，唯有如此，方能合力构建起具备超高分辨率的下一代全球气候动力学大模型。在人类文明的命运十字路口上，我实在想象不出还有什么科学议题能比查明我们是否正在跨越不可逆的气候临界点更为至关重要、迫在眉睫。只要全人类能够真正团结协作，这一科学谜题完全可以在短期内被彻底攻克。

但在更为现实的公共政策执行层面呢？假若2026年的这场旷世大旱在2027年再度原样上演（且其间冬季的降水量又未能如愿达标），英国政府眼下究竟能够紧急推出哪些切实的避险减灾政策？这个国家是否需要未雨绸缪、紧急拍板上马大规模海水淡化工厂？至少，各级政府监管职能部门必须立即就此展开跨部门的联合闭门研判。时间已经所剩无几，全人类容不得半点迟疑与懈怠。■

蒂姆·帕尔默系英国皇家学会研究数学教授、牛津大学物理系荣休教授。

ch009.xhtml

英伟达：AI领域的“中央银行”

然而，它的贷款担保真的稳健吗？

图片：卡尔·戈弗雷（Carl Godfrey）

2026年9月3日

英伟达（Nvidia）这家美国芯片巨头为全球绝大部分人工智能（AI）提供算力支持，它耗费了整整30年时间，市值才迈过1万亿美元大关；然而，从1万亿攀升至2万亿美元仅用了短短9个月；随后不到两年，其市值便突破了5万亿美元大关。如今，英伟达已成为全球市值最高的公司，估值约为5.4万亿美元。预计明年其销售额将近乎翻番，更有分析师预测，到2029年，其年营收将达到1万亿美元。

然而，推动英伟达惊人增长的并不仅限于芯片制造本身。其掌门人黄仁勋还借助财务工程手段，以此提振市场对其产品的需求。例如，在8月中旬，英伟达同意为俄亥俄州的一座大型数据中心提供高达1050亿美元的兜底担保，该数据中心将大量采用其芯片。就在一周前，英伟达还披露了一项计划：在华尔街六大投行的协助下，撬动超过5000亿美元的AI基础设施投资，其运作方式是为出售给此类项目的设备提供价值担保。在某些投资项目中，它承担的包销比例可能高达总成本的四分之一。今年7月，英伟达还以另一种方式协助部分客户筹措巨型数据中心的建设资金——承诺在其未达既定目标时为其补贴收入差额。

企业向自身客户提供贷款屡见不鲜：汽车制造商旗下的金融部门便是典型例证。黄仁勋辩称，英伟达只是在协助那些完全可行的项目释放投资潜力，否则这些项目可能难以以合理的价格借到足额资金。但在批评人士眼中，此类交易无不散发着20世纪90年代末互联网泡沫的气息——当时思科（Cisco）和朗讯（Lucent）等电信设备制造商向采购其设备的电信运营商提供了数十亿美元贷款。而当市场对电信服务需求未达预期时，部分运营商轰然倒闭，给思科和朗讯带来了惨重亏损。

正如分析机构海港研究伙伴公司（Seaport Research Partners）的杰伊·戈德堡（Jay Goldberg）所言，英伟达正在“释放既有需求”与“凭空创造需求”之间的微妙界线上走钢丝。尽管他认为英伟达目前“已相当接近”，但尚未越界。其他人士则持更加怀疑的态度，其中包括知名投资者迈克尔·伯里（Michael Burry）——他曾因做空引发2007至2009年金融危机的次级抵押贷款证券而声名大噪。伯里等人发出诘问：一旦AI芯片的需求增速低于预期，导致供给激增、价格暴跌，届时不仅会侵蚀英伟达自身的利润，更可能因其对客户的慷慨扶持而引发巨额亏损，局面又当如何收场？

庞大的“最后贷款人”

这一疑问绝非杞人忧天，因为英伟达的财务承诺数额庞大且增长迅猛。在过去三年里，它已承诺向初创企业投资逾700亿美元，并向客户提供了3000亿美元的财务支持。鉴于其在整个行业融资中扮演着举足轻重的角色，部分人已开始将英伟达戏称为“AI领域的中央银行”。正如黄仁勋所言，这固然有助于推动行业增长，但其中潜藏的风险同样不容忽视。

英伟达之所以采取财务工程手段，部分原因是其最大的一批客户正逐渐转变为竞争对手。亚马逊（Amazon）、谷歌（Google）、Meta和微软（Microsoft）等“超大规模云计算服务商”（Hyperscalers），贡献了英伟达约半数的营业收入。预计今年这些科技巨头的投资规模将达到8000亿美元左右，且绝大部分投向了AI基础设施。然而，它们当中的大多数已开始自行设计芯片，这使其未来对英伟达芯片的采购前景蒙上了一层阴影。

对于超大规模服务商而言，这些定制芯片要便宜得多，其成本仅为英伟达芯片的五分之一到三分之一。此外，得益于针对自身软件的专项优化，定制芯片在处理某些特定任务时表现更为出色。对英伟达而言更糟糕的是，少数科技巨头不仅自研自用，还开始向第三方兜售。例如，谷歌已向大型独立AI实验室Anthropic出售了部分专用处理器；亚马逊同样预计其定制芯片业务将成为相当可观的收入来源。数据服务商彭博行业研究（Bloomberg Intelligence）预测，定制芯片将逐步蚕食英伟达的市场份额，到本年代末，其在AI处理器市场的份额将从今年的约40%攀升至50%左右。

超大规模服务商拥有投资级信用评级，这使它们能够保持较低的借贷成本。相比之下，新崛起的“新型云服务商”（Neoclouds）虽有类似的大笔开支需求，营收规模却微乎其微（参见商业版块）。它们的借贷成本自然要昂贵得多。谷歌母公司Alphabet于去年11月发售了27.5亿美元的50年期债券，年利率为5.7%；而最大的新型云服务商CoreWeave在今年7月借入26亿美元时，承担的利率几乎是前者的两倍。

“英伟达银行”正是意图弥合超大规模服务商与其余厂商之间在借贷成本上的这一鸿沟。其策略之一，是入股那些未来自身会成为客户、或能间接刺激英伟达芯片需求的初创企业。去年，英伟达进行了约90笔此类投资，数量几乎是两年前的两倍；今年以来，它已敲定了另外60余笔交易。

质化宽松

英伟达开出的部分支票旨在推动开放权重（open-weight）AI模型的传播普及。与Anthropic、OpenAI和谷歌等公司通常依赖付费订阅且技术内核封闭的专有模型不同，开放权重模型允许用户免费下载和改造调整。今年8月，英伟达同意向构建AI编程模型的初创企业Poolside支付60亿美元软件授权费，并追加10亿美元入股。另据报道，英伟达已同意以129亿美元收购托管开放权重模型的开源平台Hugging Face。此类投资背后的真实意图，是通过扶植大量独立于超大规模服务商之外的AI产品与企业，持续激发市场对英伟达芯片的刚性需求。

然而，英伟达如今已不再局限于单纯投资有前景的AI公司，而是愈发直接地协助它们为巨型工程融资。今年7月初，它宣布了一项新策略：承诺为新型云服务商新建数据中心的收入提供保底支持，确保其不低于约定的底线。这些担保的具体条款因项目而异，通常为期六年。在整个期限内，英伟达承诺以固定价格购买所谓的“算力”（compute）。若新型云服务商能够以更高价格出售相关算力，英伟达则可按比例从中分成。这种安全网机制大幅提升了新型云服务商未来收入的确定性，进而压低了它们举债建设新数据中心的资金成本。而这反过来又直接拉动了市场对英伟达处理器的采购需求。

澳大利亚新型云服务商Sharon AI预计将利用此类总值约4.9亿美元的兜底担保，部署约4万张英伟达芯片；另一家新型云服务商Firmus则计划采购多达17万张。咨询机构SemiAnalysis的分析师丹·尼什鲍尔（Dan Nishball）认为，此类安排未见得会成为行业长效机制，其核心在于“争取时间”，让放贷机构逐步对向此类项目发放贷款建立信心。

以计划出租AI算力的新型云服务商CoreWeave为例。

英伟达已向CoreWeave投资逾20亿美元，持有其约11%的股份。

CoreWeave向英伟达采购AI芯片以建设数据中心。

英伟达同意在2032年前向CoreWeave采购至多63亿美元的闲置数据中心算力。

这是英伟达开展的众多采用复杂财务架构的交易之一。

它已与自身的众多客户在财务上深度绑定：包括新型云服务商、AI实验室乃至投资机构。

英伟达持有其中许多公司的股权。

它还向许多客户提供了财务兜底。理论上，这些财务承诺可能使其面临近3000亿美元的潜在成本风险。

对于算力需求庞大却处于严重失血状态的大型AI实验室，英伟达同样采取了别出心裁的策略。其在俄亥俄州支持建设的超大型数据中心归日本软银集团（SoftBank）旗下的SB Energy所有，OpenAI则作为租户入驻。英伟达将在土地和建筑物租赁合同以及项目所依赖的购电协议中充当担保人。作为交换，该项目将使用150万个英伟达处理器，并预计在项目运营周期内经历多轮“升级换代”。

即便是从亚马逊和谷歌采购芯片的Anthropic，也未能跳出这张大网。据报道，通过一系列复杂的协议安排，Anthropic签署了一项价值350亿美元的合同，从英伟达参股的新型云服务商Lambda处租赁云计算算力；而Lambda所使用的设施，又由另一家新型云服务商Hut 8开发建设，且该设施的全部容量此前已由英伟达全盘租下（目前大部分可能已转租）。

英伟达还在积极探索创新途径，以吸引更多外部资本参与此类项目。它与华尔街大行达成的5000亿美元合作计划，将瞄准主权财富基金、保险公司和养老基金等机构投资者。这些投资者将出资组建独立实体（SPV），专门用于采购英伟达硬件、构建基础设施并对外出售算力。在此类交易中，英伟达既不需要前期投入现金，也无需承担任何负债。但在某些情况下，它将提供最高达交易总价四分之一的“残值保障”（residual-value support），以弥补特定期限后相关硬件出现贬值的差额。此外，英伟达还将提供全套技术支持，使放贷方对所投资设备的充分利用更具信心。黄仁勋将这种做法定性为把算力打造成“一种可投资的资产类别”。

支撑这套繁复债务担保体系的，是两大根本假设：英伟达的芯片将持续保值，以及市场对算力的需求将继续保持飞速增长。然而，这两点均非万无一失。

图片：卡尔·戈弗雷（Carl Godfrey）

首先看芯片本身。黄仁勋坚称其芯片具备“可替代性”（fungible）与高耐久度，在售出多年后仍可通过软件升级持续提升效用。他认为，这些芯片应当被视为可靠的银行资产，足以充当贷款和担保的抵押品。然而，AI芯片的保值期限究竟有多长，业内向来争议不断。迈克尔·伯里此前便指出，大型云服务商将芯片折旧年限拉长至五到六年，而非更为切合实际的两到三年，实质上人为虚增了利润。

就目前而言，黄仁勋在辩论中尚占上风，老旧芯片依然抢手。今年8月，CoreWeave宣布签署了一份涉及英伟达A100芯片的合同，执行期一直延续至2029年。A100早于2020年便已推出，此后英伟达已陆续迭代了多款新芯片。与此同时，亦有证据表明老旧芯片维持了其内在价值：SemiAnalysis测算，按一年期合同计算，英伟达作为AI行业主力基石的H100芯片租赁价格约为每小时2.80美元，这仅比该芯片在2023年初刚推出时的价格低了约十分之一。

然而，这些芯片的寿命与保值属性，很可能只是供给稀缺下的假象。当算力供给受限时，企业别无选择，只能让老芯片继续超负荷运转。许多企业目前将此类旧芯片转用于“推理”（inference）任务——即AI模型响应用户查询的生成过程；相比耗费顶级最新硬件的模型训练而言，推理对硬件的要求要宽松得多。

英伟达自身每年都会推出性能炸裂的新一代芯片，其明确意图正是要将上一代产品迭代淘汰。这就使其陷入了一种怪异的自我矛盾境地：一方面极力主张旧芯片应保持其保值属性，另一方面又不断力推新芯片迅速取而代之。要在这两者之间取得平衡，绝非易事。

失序的前瞻指引

随着供给增加，算力价格理应逐步回落。鉴于推理开支已超越训练开支，超大规模服务商甚至OpenAI等头部实验室都在紧锣密鼓地推出自研的推理专用芯片。英伟达目前享有约75%的惊人毛利率，相比之下其体量较小的竞争对手超威半导体（AMD）仅有约55%。曾创立某AI硬件公司并随后被芯片巨头高通（Qualcomm）收购的蒂姆·戴维斯（Tim Davis）认为，随着市场可选项的激增，芯片制造领域的暴利结构“将开始承压收缩”。

对芯片价格构成的最大威胁，归根结底仍是需求本身。只要AI领域的资本开支保持繁荣，英伟达就能顺利售出芯片，其客户的数据中心也能被算力需求填满，而其作出的各项兜底担保便极少被触发。然而，一旦需求扩张不及市场乐观预期，新型云服务商和其他AI供应商就可能难以售罄算力或实现盈利。这进而可能触发英伟达的兜底条款，迫使其自掏腰包购买闲置算力、补足价格缺口或承担无端闲置的电力费用。这种需求放缓同样会重创英伟达自身的销售业绩，从而削弱其用于履行这些兜底义务的自由现金流。这种险境甚至无需以需求崩溃为前提——仅仅是低于预期的增速疲软，就足以在业内掀起惊涛骇浪。

那么，英伟达究竟可能需要自掏多少真金白银？它目前已承诺了约250亿美元的未来股权投资，自身背负着约330亿美元的债务。而其对客户承担的潜在负债规模约为3000亿美元，但由于此类负债只有在行业下行期才会生效，因而并未反映在当下的资产负债表内。其中包括为OpenAI数据中心提供的1050亿美元担保、通过华尔街合作项目提供的多达1250亿美元担保，以及约670亿美元的其他兜底承诺。

英伟达最终几乎绝无可能支付接近其担保总额的资金。它的财务承诺分散在未来数年之内，绝不会在一夕之间全数到期。例如，其拟为OpenAI数据中心提供的兜底担保从2028年开始生效，期限长达20年。倘若OpenAI届时违约停付租金，英伟达亦可另寻租客接盘，从而大幅缩减自身面临的实际敞口。

此外，英伟达自身的财务实力也足够强劲，足以抵御此类潜在负债的冲击。投资银行摩根士丹利（Morgan Stanley）预计，随着各项担保逐步生效，英伟达的“全包式”负债总额将从明年初的530亿美元飙升至2029年初的2000亿美元。然而，这足以被其目前价值990亿美元的充沛现金及高流动性证券所对冲，加之其主营业务今年将产生约2000亿美元的强劲现金流。就当前格局而言，除非遭遇毁灭性的全面崩盘——导致英伟达的所有担保同时到期且公司利润近乎彻底化为乌有，否则它很难面临真正的生存危机。

然而，若英伟达的表外承诺规模继续无休止膨胀，局面或将完全不同。金融研究机构CreditSights的安迪·李（Andy Li）担忧，英伟达可能会一路“猛踩油门”，直至“车毁人亡”。SemiAnalysis测算，在其新型云服务商兜底计划涵盖的数据中心容量中，每100兆瓦（MW）就对应着英伟达约59亿美元的担保规模。SemiAnalysis预计，若英伟达继续追加此类担保，仅该计划一项，到2028年底其敞口规模就可能达到1750亿美元。

在利用日益复杂的金融衍生工具来刺激客户需求方面，英伟达并非孤例。AMD此前便提出向OpenAI和Meta出售大量自身股权，以换取对方采购其芯片的巨额合同。今年4月，谷歌牵头组建了一家企业财团，协助Anthropic筹措用于购买价值350亿美元谷歌芯片的资金。代表谷歌开发相关芯片的博通（Broadcom）则同意在Anthropic无力偿付时代为填补资金缺口。

在这场竞相锁定销售额与贷款规模的狂热竞逐中，潜藏着一个巨大风险：那些在正常市场环境下根本难以获得融资的投机性项目，如今正被盲目推向建设。一旦行业的实际回报不如预期，或是兑现周期远长于这些盘根错节的交易所预设的时间表，最终留下的将是铺天盖地的昂贵闲置芯片。在这张错综复杂的金融巨网中，没有哪家企业陷得比英伟达更深。黄仁勋坚信，当前AI基础设施的建设浪潮正“全速推进”。但英伟达的投资者们必须保持清醒：在这场繁荣盛宴中英伟达垫付的资金越多，一旦幻象破灭，它所需咽下的苦果也将最为深重。■

ch010.xhtml

钱德拉巴布·奈杜为其邦构想的科技乌托邦蓝图

这位精通科技的安得拉邦首席部长是否已好高骛远、不堪重负？

摄影：迭戈·马洛（Diego Mallo）

2026年9月3日 | 阿马拉瓦蒂

每当讲起这段往事，钱德拉巴布·奈杜总是面带微笑。1995年出任印度东海岸安得拉邦首席部长时，他敏锐地察觉到“信息技术（IT）就是未来”，随即启程前往美国。起初比尔·盖茨拒绝接见他，但奈杜态度坚决。到了1998年，微软舍弃了班加罗尔，选择将印度总部设在该邦当时的首府海得拉巴。如今，海得拉巴拥有上万名微软员工、一座大型国际机场，城市人口翻了一番。在奈杜眼中，盖茨不仅是一位投资者，更是推动海得拉巴崛起的一座引路灯塔。

这个故事生动勾勒出了奈杜的公众形象——一个目标坚定、精通招商引资之道并矢志革新印度治理水平的科技远见者。历经政海沉浮，如今76岁的他依旧雄心勃勃，已进入其第四个任期的第二年。他曾于2003年险遭刺杀，在2023年短暂身陷囹圄，并曾两度被选民赶下台。最令他痛心的是，2014年他的邦被一分为二：核心掌上明珠海得拉巴被划为新成立的特伦甘纳邦首府，而留在安得拉邦的，只剩下一片面积堪比突尼斯、人口达5400万的农业残缺板块，既无首府，亦无经济引擎。

地图：《经济学人》

一项令人眩晕的宏图大计由此诞生。作为这一残缺之邦的首席部长，奈杜规划了一座名为“阿马拉瓦蒂”的璀璨新首府——这是印度历史上规模最大的全新规划建设城市（greenfield city），也是全球现阶段在建体量最大的新城之一。他还为安得拉邦确立了全新的发展定位：不再与印度原有的软件中心竞争，而是将其打造成先进制造业、数据中心和量子计算的核心聚集地。为实现这一宏愿，他率领其地方政党与德里执政的纳伦德拉·莫迪领导的印度人民党（BJP）结盟，作为联合执政伙伴换取中央的财政鼎力支持。

阿马拉瓦蒂的建设工地上，2.5万名工人昼夜不休地在3.3万英亩土地上忙碌，全力将蓝图化为现实。在紧邻工地的办公室内，奈杜向《经济学人》表示，他正在打造全印度最美丽的城市。这一借鉴新加坡经验的总体规划环绕网格化系统展开，拥有丰富的水系、丰沛的绿化和绵延数英里的自行车道。他预测，五年内大街上就将行驶着自动驾驶汽车。目前该地已划拨约300英亩土地专门用于无人机的研发与制造，他设想新城在不久后每年将迎来1万架次无人机飞行，让居民能够漫步在绿树成荫、毫无大型货车喧嚣的林荫道上。

奈杜的蓝图听上去宛如乌托邦，但他坚称这植根于现实。长期以来印度在硬件领域始终落后，但目前正逐步崛起为电子制造业基地。在交谈中，奈杜频频提及中国的制造业实力，认为印度必须引为效仿。在他看来，安得拉邦拔地而起的工厂、数据中心和量子计算机，将与班加罗尔和海得拉巴的初创企业及外包枢纽形成互补，而非直接竞争。

然而，眼前的重任极为艰巨。1995年时的海得拉巴已具备科技产业底蕴并坐拥500万人口；而今，奈杜却试图在几片老旧的香蕉园地基上，从零构筑一座高科技经济中心。更为严峻的是，该项目争议缠身，邦政府财政状况捉襟见肘。2019年，奈杜在选举中败给了一味推行民粹福利主义的对手政党YSR大会党（YSR Congress）。后者将阿马拉瓦蒂项目搁置冷冻长达五年。为了在2024年夺回政权，奈杜不得不承诺保留大量成本高昂的福利计划，这进一步挤压了可用于基建投资的财政资源。

摄影：.

尽管如此，他依然取得了亮眼的政绩。在港口城市维沙卡帕特南，谷歌于今年4月启动了一项1吉瓦（GW）的AI数据中心工程，这是全印度规模最大的同类项目。这一耗资高达150亿美元的合资项目拉拢了莫迪的亲密盟友、印度最具权势的商业大亨之一高塔姆·阿达尼（Gautam Adani）参与其中，充分印证了奈杜老辣的政治手腕。此外，他大力优化营商环境的举措也起到了关键作用，钢铁、国防和航空航天类企业如今也在纷至沓来。

在阿马拉瓦蒂，奈杜再次施展拳脚引进行业标杆企业——包括科技巨头IBM，期望借此引发群聚效应。目前已有三所大学和约10万名学生穿梭生活在林立的塔吊与挖掘机之间。在原本的沼泽地上，宏伟公共建筑的雏形已然拔地而起。全城共有95个地块正同步开工。“这里未来要放置超级计算机！”一位工头在机器轰鸣的喧嚣声中兴奋地喊道。

在这个饱受平庸官僚困扰的国家，奈杜的雷厉风行令人耳目一新。即便如此，他可能依然有些好高骛远、力不从心。批评者指出，阿马拉瓦蒂是一项野心过大、劳民伤财的面子工程，很可能沦为拖垮全邦财政的沉重枷锁。政治经济学家帕拉卡拉·普拉巴卡尔（Parakala Prabhakar）表示：“他这是在蚍蜉撼树、不自量力。”该市主要的建设资金来源于世界银行（World Bank）和亚洲开发银行（ADB）提供的16亿美元长期贷款，未来需通过出让土地来逐步偿还。这种账面逻辑完全依赖地价暴涨，而这又反过来取决于新城能否吸引到足够的外国投资——可偏偏眼下全印度在吸引外资方面都正处于艰难期。

“这些项目在昂贵的3D渲染效果图里向来宛如天堂，”麦吉尔大学（McGill University）学者莎拉·莫瑟（Sarah Moser）评论道。她曾建立了一个涵盖全球300座新建规划城市的数据库，“但其中绝大多数最终都半数荒废，有的甚至沦为无人鬼城。不论奠基人多么有远见卓识，他们往往都难以扭转经济地理规律的客观现实。更何况，政治风向瞬息万变。不妨看看沙特沙漠中那座庞大的阿卜杜拉国王经济城，自其同名开创者于2015年辞世后，至今仍在黄沙中积灰蒙尘。”

阿马拉瓦蒂的未来命悬于脆弱的政治平衡之上。一旦奈杜失势下台，继任对手必然会另立主张。目前，他之所以有底气强行推进工程，是因为人民党虽在全国执政，却未能取得议会绝对多数席位，不得不倚赖其盟友地位。然而这种格局随时可能生变。奈杜声称他将与莫迪长期风雨同舟，因为双方在发展理念上“步调一致”（且反对党联盟依旧“脱离实际”）。但他亦坦承，近期年轻人群体爆发的抗议活动是“摆在所有人面前的严峻挑战”。

最后一个悬而未决的疑虑，在于阿马拉瓦蒂能否真正跃升为“量子谷”。尽管当前资金如潮水般涌向人工智能，但许多行业专家认为，量子计算机未来将发挥互补作用，部分原因在于其在执行某些特定任务时的能效将远超大语言模型。奈杜希望阿马拉瓦蒂能承担相关零部件的制造。然而，这项技术距离真正商业化落地尚有数年之遥，究竟哪种量子技术路径能够最终胜出也仍未可知（IBM专攻其中一条技术路线，这带来了被单一技术锁定的沉没风险）。尽管布局量子计算或许是一场深谋远虑的长期豪赌，但在短期内显然难以迅速创造投资与就业岗位。

正如当年打造海得拉巴一样，奈杜坚信这片残缺的土地能够通过尽早抢占新兴技术先机而实现腾飞。展望未来，他希望将权力交接给目前担任邦科技部长的儿子纳拉·洛克什（Nara Lokesh）。但他目前毫无退隐之意。“人，”他说，“永远都要往前看。”■

ch011.xhtml

随着中国展现外交肌肉，太平洋岛国论坛惨淡收场

关于台湾出席资格的纷争导致多国领导人缺席

惠普斯巡视一圈 © 摄影：盖蒂图片社（Getty Images）

2026年9月3日

太平洋岛国论坛（PIF）这个由18个成员组成的区域性组织有一项独特传统：在领导人峰会期间，夏威夷花衬衫几乎是“必备着装”。在8月30日至9月1日举行的本届会议上，无论是太平洋岛国领导人，还是来自美欧及亚洲其他地区的高级外交官，均着此类服饰亮相。

然而，另一项不那么轻松的现实是：该论坛有三个成员国与台湾而非中国大陆保持着外交关系。当其中某个邦交国轮值主办年度峰会时，便会出现罕见的一幕——台湾最高级别外交事务官员林佳龙得以像正规外长一样穿梭游走、举办新闻发布会。

地图：《经济学人》

今年，各国领导人齐聚帕劳——全球仅存的11个承认台湾的国家之一。中国大陆的愤怒可想而知。但其为扰乱会议进程所采取行动的深度，依然出乎外界预料。五位领导人在最后一刻临时退场，给论坛带来了沉重打击。太平洋岛国论坛的决策机制严格遵循协商一致原则，且唯有国家领导人亲自到场方能敲定关键重大决议。

外界很快将怀疑的目光投向中国大陆。帕劳总统苏兰格尔·惠普斯（Surangel Whipps）长期批评中国大陆在该地区战略要地岛国收买影响力的行径，他向记者透露，早在数周前，中国大陆就曾发出警告，称如果帕劳邀请台湾，将面临“难以预料的后果”。

缺席名单中最引人瞩目的是自今年5月中旬起出任所罗门群岛总理的马修·威尔（Matthew Wale）。作为论坛的卸任主席，威尔此前一直在积极游说各方通过一项新条约，该条约旨在提高中国大陆在论坛成员国建立军事或警务驻留的门槛。然而就在峰会开幕前夕，他的三名内阁部长突然请辞，并对其发起了不信任动议。威尔不得不匆忙赶回国内平息议会倒阁危机，导致他在帕劳倡导的条约草案彻底失去了领头人。

所罗门群岛首都霍尼亚拉的外交人士透露，他们经常看到中国大使宴请该国议员，尤其在威尔出任总理后更是如此。一位外交官称，中国外交官与国有企业频繁向议员分发资金，以诱使其支持对华更友好的政策。不过，他们并未能拿出确凿证据，证明这几名部长的突然反水直接受控于中国大陆在此次事件中的暗中运作。

部分领导人的缺席或许亦有其他缘由：有些领导人身体抱恙，有的则分身乏术于国内政局。然而，另一位临阵退缩者——基里巴斯总统塔内蒂·马茂（Taneti Maamau）的举动却毫不掩饰，他竟直接指派该国前驻华大使代为出席。这一切令台湾的林佳龙大为光火，他向记者表示，中国大陆宛如一个喧宾夺主的客人，四处颐指气使，“彻底搞砸了聚会的气氛”。■

ch012.xhtml

灾难性洪灾警示尼泊尔：更多巨灾接踵而至

该国极度匮乏应对灾难所需的资源与国际援助

无家可归 © 摄影：美联社（AP Photo）/拉杰什·库马尔·辛格（Rajesh Kumar Singh）

2026年9月3日 | 德里与加德满都

在喜马拉雅山区遭遇这场浩劫整整一周后，8月26日发生的那场天灾所造成的破坏广度仍难以被全面评估。不过，人们至少已能相当精准地还原其爆发的起点：上午8时37分，伴随着冰川崩塌，一场巨大的冰岩崩塌倾泻而下，截断了尼泊尔与中国西藏交界处的楞德河（Lhende river）。仅仅三分钟后，形成的堰塞湖溃决，裹挟着巨量泥石、残冰的狂暴洪流顺着尼泊尔热索瓦（Rasuwa）地区的波特科西河（Bhote Koshi）与特耳苏里河（Trishuli）倾泻而下。目击者希布·吉里（Shibu Giri）眼睁睁看着滔天巨浪将一整座桥梁夷为平地；就在巨浪将他的卡车卷走前的一刹那，他拼死逃入了密林。在山洞中熬过一夜后，他于次日被军方直升机营救脱险。然而许多人并未如此幸运，多座村庄在顷刻间被山洪彻底吞没。

地图：《经济学人》

一周后，已确认的遇难人数超过1000人，失踪人数逾4000人。罹难者中包含数百名外国游客，其中包括多辆大巴上前往西藏神山冈仁波齐峰朝圣的印度教信徒。灾区约80座桥梁被毁，40公里道路化为乌有；14座水力发电站遭到重创（部分尚在建设中），直接导致尼泊尔全国一成的发电产能瘫痪。

尼泊尔地处印度洋板块与欧亚板块剧烈碰撞挤压的地带，炽热的季风水汽在此迎头遭遇喜马拉雅山脉的刺骨寒风与脆弱山体。这一独特的地理构造使其成为全球地质与气象灾害最频发的地区之一。该地区的升温速度远超地球其他区域，冰川消融速率相比2000年已翻了一番。这意味着在未来的岁月里，势必将有更多冰川解体崩塌，更多冰川湖面临决堤之虞。

作为全球最贫困的国家之一，尼泊尔在应对接踵而至的极端灾害面前显得极度脆弱无力。该国总理巴伦德拉·沙阿（Balendra Shah）现年36岁，曾是一名说唱歌手。一年前尼泊尔青年焚毁议会大厦并推翻前任政府后，他掌权至今不过短短六个月。如今，其政府必须筹措高达50亿美元的灾后重建资金，而以往尼泊尔赖以求援的富裕国家，此刻却纷纷在削减对外援助预算。

联合国气候基金甚至已无力满足贫困国家现有提出的援助申请。美国已承诺提供360万美元的救灾援助——尚不足其在2015年尼泊尔大地震后援助金额的3%。资金的极度短缺，不仅极大制约了当下的紧急人道主义救援行动，更令通过投资建设具备气候抗逆性的道路、桥梁及防灾村落以实现“重建得更好”的宏伟规划彻底沦为空谈。

与此同时，喜马拉雅沿线各国彼此间戒心重重、锱铢必较，难以携手协调应对这一共同家园的生态危机。兴都库什-喜马拉雅地区拥有超过6.37万条冰川，却仅设有38座监测站（其中符合国际监测标准的仅有区区7座）。设在尼泊尔首都加德满都的研究与政策机构“国际山地综合发展中心”（ICIMOD）总干事白玛·嘉措（Pema Gyamtsho）指出，尼泊尔必须仰仗资源更为充裕的邻国，为该区域布设更先进的传感器与监测设备，并实现水文数据的实时共享。然而，中国往往将跨境水文流量数据视作国家安全机密，使得下游邻国时刻面临沦于被动盲区的重大隐患。

面对如此突如其来且势不可挡的狂暴山洪，或许即便有早期预警系统，也未必能为尼泊尔争取到充裕的避险筹备时间。然而，要想在未来数年由区域山川水系酿成的巨灾中幸存下来，尼泊尔迫切需要更多的资金注入，以及更高程度的跨境透明合作。■

ch013.xhtml

日本主流在野党阵营陷入一片混乱

新联盟的破裂折射出更为深层的内部顽疾

摄影：盖蒂图片社（Getty Images）

2026年9月3日 | 东京

“中间派改革联盟”（CRA）曾怀揣着宏大抱负粉墨登场。今年1月，日本两大主流在野党宣布在大选前合并其在握有实权的众议院内的党团力量，意图与执政的自由民主党（LDP，简称自民党）一较高下。然而结局却是自民党大获全胜，中间派改革联盟惨遭滑铁卢——其在此前持有的167个席位中，最终仅艰难斩获49席。8月28日，该联盟成员宣布未能在参议院就合并事宜达成共识；仅三天后，该联盟便宣布在众议院同样就地解散，距离其初创尚不足一年。

中间派改革联盟的失败，暴露出日本主流在野党阵营根深蒂固的顽疾。该联盟融合了曾于2009至2012年掌权的在野党的主要衣钵继承者——中左翼的立宪民主党（CDP），以及曾长期与自民党组建执政联盟的公明党（Komeito）。然而两党皆受制于后继无人、难以推举出令选民耳目一新的年轻领袖，在争取青年选民方面更显得束手无策。

这场狼狈的溃败意味着自民党面临的制衡压力大幅减轻，也为新兴民粹主义政党腾出了扩张空间。自民党自1955年创立以来便长年垄断日本政坛。当日本民主党（DPJ）在2009年横扫大选强势上台时，曾被寄予开启两党轮流执政政治新纪元的厚望。但民主党的执政过程却混乱不堪；更不幸的是，它在任内赶上了2011年的东日本大地震与海啸灾难，进而引发了福岛核电站严重堆芯熔毁事故。这场核灾的政治余波至今仍未消散，使作为其法定继承者的立宪民主党彻底沦为选民避之不及的“政治放射源”。

近年来，频发的政治献金黑金丑闻虽令自民党声誉扫地，使其在2024年和2025年的选举中接连丢失大量席位，公明党甚至在去年10月愤而退出长达数十年的执政联盟，以抗议新任首相高市早苗治下自民党政策路线的急剧右转。立宪民主党与公明党本期望通过合流重夺政治中间路线的主导权。据报道，两党甚至曾试图策动高市早苗的前任、来自自民党内部自由派阵营的两位前首相倒戈加盟。

然而，无论面对潜在的政治盟友还是普通选民，它们的政治诉求均未激起半点回响。为了强行凑在一起，两党不得不在各自的核心政治原则上大肆妥协。加之身为首位女性首相的高市早苗展现出充沛的政治活力，反衬得中间派改革联盟由两位耄耋男性把持的领导层宛如步履蹒跚的恐龙。一位立宪民主党议员沮丧地抱怨道：在社交媒体时代，这群党内元老本应“主动退居二线，让更年轻的面孔站在台前”。

大选惨败后，该联盟迟迟未能厘清自身的施政纲领。绝大多数选民普遍认为，两党本就该各走各路、分道扬镳。两党合并谈判的最终破裂完全在情理之中。东京上智大学学者中野晃一（Nakano Koichi）直言，谁都不想跳上一艘正在沉没的破船，尤其是这艘船本身还如此微不足道：“这甚至都算不上是一艘‘泰坦尼克号’——它不过是一叶无关紧要的偏狭扁舟罢了。”■

ch014.xhtml

印度国家支付系统迎来十周年：必须学会自负盈亏

月处理240亿笔交易的高昂成本绝非从天而降

插画：莱赫尔·科瓦奇（Lehel Kovács）

2026年8月30日

在孟买海滩货摊的椰子顶端、德里街边小吃店的萨莫萨三角炸咖喱角旁、加尔各答烟摊的“海军切牌”（Navy Cut）香烟盒间，处处可见它的身影。它被张贴在人力三轮车司机的椅背上、公用事业公司的账单上，乃至亚马逊平台的结账界面中。无论是买10卢比（约合10美分）的花生，还是购买价值10万卢比的iPhone，只需扫它一下即可完成支付。如今，这个二维码已无所不在、深入寻常巷陌，以至于人们往往容易遗忘，它融入印度社会风貌其实不过是弹指一挥间的事情。

这个二维码正是印度“统一支付接口”（UPI）的标志性视觉符号。这一平台于8月25日迎来了其十周岁生日，它彻底颠覆了印度资金流转的运作模式。若回溯到遥远的“旧石器时代”——也就是2016年之前，以交易笔数计，现金几乎承载了全印度所有的交易；即便按交易金额算，现金也占据了整整三分之二。而在非现金交易中，支票是绝对的主流。不仅如此，那时几乎所有人都会出于某种心照不宣的原因坚称自己“没有零钱”，由此甚至催生出了一种将1卢比的水果硬糖充当法定找零替代货币的独特奇景。

在UPI上线后的前12个月里，全印度民众通过该平台累计完成了7300万笔交易，交易额达2270亿卢比。而到了今年，仅8月份单月，该系统的交易次数就高达250亿笔，流转资金近30万亿卢比。目前，印度超过85%的非现金交易均通过UPI完成（而支票结算的占比已断崖式跌至区区0.2%）。更重要的是，其增长潜力仍未见顶：该网络拥有的约5.5亿活跃用户，仅占印度成年人口总数的一半出头。

UPI之所以能取得如此瞩目的成功，原因不一而足，而首当其冲的当属极致的简便性。用户只需将其银行账户与手机应用程序绑定，即可畅通无阻地收付款项。自此之后，完成支付就像扫描二维码或输入手机号码一样轻松惬意——无需披露敏感的银行账户信息，不必繁琐地翻找塑料银行卡，更从此彻底告别了线下POS机。

多重时代红利的叠加共振亦功不可没。印度电信运营商Jio与UPI于同年横空出世，凭借其极具杀伤力的激进低价策略，促使移动流量资费呈现断崖式暴跌，将数以亿计的印度平民接入了互联网世界——进而将他们引流至UPI平台。此外，莫迪政府在同年11月出人意料地祭出雷霆手段，以打击隐匿黑钱为由宣布作废占当时流通现金总值86%的500及1000面额卢比纸币。这一“废钞令”政策虽重创了非正规经济并导致GDP增速显著放缓，却不出所料地催化了无现金支付的爆发式普及。UPI的月交易量瞬间从2016年10月的10万笔，雪崩式暴涨至同年12月的200万笔。

另一项起到助推作用且负面效应小得多的政策，是政府明令禁止对UPI商业交易征收交易手续费。在传统模式下，商家接受信用卡或借记卡付款通常需要支付一笔刷卡费，该费用由网络清算机构（如Visa或万事达卡）、发卡行等多方瓜分。然而自2020年起，印度商家通过UPI收款无需缴纳分文费用，这进一步掀起了商户争相接入的狂潮。

就在UPI迎来十周岁生日的前夕，印度议会表决通过了废止这道商户“免佣金禁令”的法案。尽管具体的费率标准尚未正式出台，但各界普遍预计新费率将仅针对商业交易，并设有一定的免征起征点，从而豁免绝大多数小额交易，包括向路边椰子小贩、三轮车夫付款或个人之间的转账。其核心构想在于“劫富济贫”，让大企业承担基础设施成本。令人啼笑皆非的是，在一向对巨头垄断深恶痛绝的印度，社会各界对这一拟议收费政策的舆论反应却以压倒性的负面为主。批评者辩称：政府完全可以通过减少印钞节省下来的财政支出冲抵系统运维成本；UPI创造的社会经济红利远超其运营开销；更有甚者断言，这笔新增成本最终依然会被转嫁到普通消费者头上。

最后一种论调听起来似乎颇有道理，但UPI的日常运转确实耗资不菲。该系统由印度国家支付公司（NPCI）负责监管运营，这是一家由公共与私营金融机构共同出资设立的非营利组织。各大商业银行为了维护庞大的网络运行、严防金融欺诈并确保每一笔支付能够稳定清算，每年需倾注天量资金。作为补偿，政府向各银行发放的财政补贴却仅能覆盖其运维成本的九牛一毛。若长期缺乏充沛的资金回血渠道，银行将毫无动力对系统升级展开创新投资；更糟糕的是，它们甚至连日常的基础设施维护和反欺诈风险防御都已捉襟见肘、难以维系。事实上，消费者早已在买单——只不过是以纳税人身份间接承担，且被迫忍受着本可以更加卓越的服务质量。

印度政府在打造UPI平台、将其树立为国家支付命脉以及两次针对手续费因势利导作出决策等方面，确实功不可没。然而，倘若这一系统因失血而陷入停滞甚至脆弱衰败，上述辉煌成就必将付诸东流。多年以来，UPI始终是一座让全社会免费享受资金流转的桥梁；如今，它自身也到了必须学会自负盈亏的时刻。■

ch015.xhtml

限制青少年使用社交媒体的禁令在亚洲难见成效

禁令遭到广泛漠视、公然践踏与花样规避

……大人还自以为掌控了一切 © 摄影：盖蒂图片社（Getty Images）

2026年9月3日

这曾看似是一剂拯救家长于水火的良方。为了管束14岁儿子耗在网上的时间，哈里妮·奈杜（Harini Naidu）使尽了浑身解数：从限制Wi-Fi接入到每晚准时拔掉路由器插头，招式用尽。因此，当澳大利亚成为全球首个立法禁止16岁以下未成年人拥有社交媒体账户的国家时，她曾满怀憧憬。然而，这项自去年12月起正式施行的法案，“在现实中根本形同虚设”。她的儿子虽然失去了YouTube账号，却依然无需登录便能随意浏览视频；而她年满15岁的女儿伊莎娜（Ishana），其在X（原Twitter）、Instagram、Snapchat和TikTok上的账号更是安然无恙、照用不误。伊莎娜直言不讳：“这政策推得软弱无力，大家早就把这道禁令抛到九霄云外了。”他们甚至根本无需动用虚拟专用网络（VPN）进行隐匿。

像伊莎娜这样的青少年的亲身体验，对全球其他国家而言同样具有标杆意义。自澳大利亚出台这道禁令以来，已有数十个国家跟进效仿或正积极酝酿类似举措。8月24日，新西兰宣布了推行类似禁令的立法意向。政策制定者普遍担忧社交媒体会对青少年心智造成全方位毒害，同时他们也深知严苛的禁令在成年选民群体中极具号召力。8月26日，Meta在就多起指控其平台损害未成年人身心健康的诉讼达成和解后，同意在美国大部分州对其面向18岁以下用户的产品功能进行全面整改，包括强制推行每日两小时的刷屏时长上限，此举赢得了社会的广泛喝彩。

然而迄今为止的事实证明，澳大利亚的铁腕政策在绝大程度上沦为了一纸空文。起初，官方还曾大肆宣扬部分阶段性战果：在法案施行的首月内，各大社交平台累计清理、封禁或冻结了近500万个可疑账户。或许眼下判定澳大利亚新规彻底折戟尚为时过早，但多项独立研究表明，新规对青少年的实际使用习惯几乎未产生实质性触动。今年6月发表在《英国医学杂志》（BMJ）上的一项权威研究指出，在限制措施落地三个月后，仍有85%的16岁以下受访者坦承自己依然在使用被明令禁止的社交平台；12岁至13岁群体的社交媒体日均在线时长与禁令出台前毫无二致。此前社会各界曾广泛担忧孩子们会因此集体转战至安全防范标准更为低劣的小众替代平台，如今看来这种担忧纯属多虑——但其核心原因，仅仅是因为绝大多数孩子依旧牢牢盘踞在原有的主流平台上。

澳大利亚的年轻一代有着极强的在线社交执念。伊莎娜透露，在许多人看来这甚至演变成了一场反叛的游戏，“正因为被禁了，大家反而更渴望玩社交媒体”。她解释道，年轻一代“太清楚该如何戏耍监管系统了”，比如与朋友创建联名共用账户，这种骚操作往往能直接把现有的年龄核验算法彻底绕晕。更具讽刺意味的是，她表示自己有好几个如实填写了真实出生日期的账号，至今也根本未被平台处置。许多16岁以下的青少年直接虚报一个假年龄重新注册账号，还有人干脆堂而皇之地借用父母或成年兄妹的身份信息登记。

不少人将这一乱象归咎于科技巨头在落实合规上消极怠工。软件测试机构KJR近日在澳大利亚境内尝试注册了50个声称用户年龄为16岁的社交媒体账号，测试结果显示仅有一家网站履行了实质性的年龄核查程序，最终所有50个测试账号全数顺利激活上线。为此，澳政府正试图通过将对违规科技企业的罚款上限翻倍至9900万澳元（约合7100万美元）等手段以加大威慑力度，并计划赋予澳大利亚电子安全委员会（eSafety Commissioner）强制调取社交媒体企业内部用户留存数据的执法权。

另一部分观察人士则认为，政策迄今难见成效，应部分归咎于支撑禁令落地的技术手段本身存在天然缺陷。各大科技平台目前正尝试借助AI文本分析和人脸扫描识别等多样化工具，但现阶段的人脸测龄软件在分辨具体年龄段时精度极为有限，例如根本无法精准区分一个人究竟是15岁还是16岁。来自维多利亚州的15岁少女露西娜（Lucina）表示，她在Snapchat进行人脸扫描验证时，系统竟误判她已有18岁。“这项政策根本没有经过实质检验，因为在技术执行层面它几乎是一道无法落地的死题，”珀斯科廷大学（Curtin University）学者塔玛·利弗（Tama Leaver）如是评价。

利弗认为，其他拟推行类似立法的国家如今正逐步达成一项共识：即“唯一切实可行的路径，是在用户注册社交媒体时强制推行法定真实身份核验（Real ID）”——即核验官方颁发的实体或数字化身份证件。作为少数几个同样推行澳式限制令的国家之一，马来西亚便义无反顾地选择了这一铁腕路径。马来西亚监管机构通讯与多媒体委员会（MCMC）首席法律顾问L.S.莱昂纳德（L.S. Leonard）表示，经其全面评估，这一方案在精准度上将“远超”澳大利亚所依赖的粗糙的年龄推定法。

然而，马来西亚的落地进展同样未能按预想推进。莱昂纳德称，各大社交平台均坚称已按照自身合规准则清理了海量未成年人违规账号。但诺丁汉大学马来西亚分校学者特里西亚·杨（Tricia Yeoh）指出，对于当地绝大多数普通用户而言，实际现状根本未发生丝毫改变。莱昂纳德对此回应称，各大平台已正式向官方申请宽限期，以便“对其系统完成改造升级”以适配身份证件核验；他预计所有技术障碍将在明年1月前“得到彻底解决”，届时仍未合规的平台将面临重罚。然而特里西亚·杨敏锐地指出，各大平台真正的顾虑，在于在自身服务器中存储海量公民个人身份核心凭证所潜藏的系统性风险——这些数据库随时面临黑客渗透窃取的梦魇，或是被迫沦为满足政府各类监管调取指令的工具。

亦有反对者认为，澳大利亚这种一刀切的粗暴封禁法令，实质上是在纵容科技平台逃避主动营造安全网络环境的主体责任。目前部分国家正在积极探索反向倒逼机制：加拿大计划立法禁止平台向16岁以下未成年人开放账户，但同时设立了豁免出口——即只要相关App通过一系列严苛的安全防护标准核验即可解禁。此外，社会各界亦寄望于Meta在美推行的防沉迷保护功能能够逐步推广至全球其他市场，进而带动更多社交巨头亦步亦趋跟进整改。

澳大利亚的年轻一代则认为，成人社会未免小觑了他们的数字素养：哈里妮·奈杜14岁的儿子坦言，同龄人中其实绝大多数都“很清楚该如何在网络世界中保护自身安全”。以他自己为例，他平日里对X平台敬而远之，因为上面的内容“缺乏底线过滤”；他反而更青睐Reddit，因为那里的社区氛围“远没有那么歇斯底里和混乱失控”。■

ch016.xhtml

科技日新月异的中国，为何管不好舌尖上的安全？

毒蔬菜成为最新食品安全丑闻的焦点

图片来源：新华社/eyevine

2026年9月3日 | 康保

这一次，让中国消费者避之不及的是大白菜。8月22日，一名博主发布了视频，曝光在距北京仅数小时车程的康保县，有货车司机将大白菜浸入甲醛溶液中。这种化学试剂能在炎炎夏日的长途运输中让菜叶保持鲜绿，但甲醛同时也是一种致癌物。在举国公愤之下，当地政府设立了检查站，对运出该县的大白菜进行检测。截至目前，已有三人因涉嫌刑事犯罪被拘留。新加坡和韩国也加强了对进口中国大白菜的检验检疫。

官方媒体痛斥涉事货车司机“突破道德底线”。然而，许多中国民众早已见怪不怪，只感到深深的无奈。康保当地一名农工承认，他早就知道这一情况。有关滥用甲醛的报道至少可以追溯到2012年。而在中国屡禁不止的各类有害食品添加物中，甲醛不过是冰山一角。就在上周，一家清洁公司因在餐馆使用违禁农药而接受调查；近期还有杨梅商贩被抓到用另一种违禁化学制剂浸泡杨梅以保鲜。

在中国，鲜有事情能像食品安全丑闻那样激起公愤。许多人至今仍对2008年的“三聚氰胺毒奶粉”事件心有余悸，当年那起事件导致约30万名婴幼儿患病，至少6人死亡。2013年，中国最高领导人习近平将食品安全问题与执政合法性联系在一起。他指出：“如果我们党在全国执政，连个食品安全都做不好……那是一定要被问到是不是合格的。”然而，丑闻依然层出不穷。随着国家日益强大和现代化，这一问题反而愈发令人尴尬。人们可以买到世界领先的国产机器人、电动汽车和智能手机，却依然无法放心食用端上餐桌的食物。

相关整改并非毫无进展。中国对2009年出台的全面《食品安全法》进行了多次修订完善。当前的监管法规严格程度并不逊色于欧洲。与此同时，处罚力度往往要严厉得多：三聚氰胺奶粉案中有两名主犯被执行死刑。中国在农药治理方面也取得了实质成效。策纬咨询（Trivium China）分析师裴埃文（Even Pay）指出，得益于针对农户的培训、农作物检测以及对高毒农药的严厉禁令，农民如今使用的农药量已大幅减少，这意味着端上餐桌的有害农残随之降低。一项针对中国西南部四川省的研究显示，从2010年到2020年，当地亩均农药使用量下降了三分之一。

然而，挑战的庞大规模依然令人望而生畏。在任何国家，确保食品安全都绝非易事：今年夏天，美国也在应对一场因生菜引发的严重寄生虫感染疫情。而在中国，食品行业的高度分散让监管难上加难。少数几家大型农业龙头企业，面对的是1.6亿散户农民，后者大多在欠发达地区的零碎土地上艰难谋生。此外，还有数以万计的中小物流商负责将粮食、果蔬从田间地头转运到城市。大多数中国百姓日常买菜依然习惯去社区周边的菜市场，而非大型连锁超市；许多人则频繁光顾街头巷尾的小餐馆。这一切为不法分子留下了巨大的操作空间，微薄的利润率更促使他们铤而走险、偷工减料。

监管体制的碎片化程度与行业本身不相上下，导致监管真空屡见不鲜。2024年，媒体曝光了散装食用油在运输过程中与化工用油混用罐车、且换货底不清洗的惊人内幕。遏制此类违法乱象本应是市场监管部门的职责，但也可能落在交通运输部，或是负责石油等战略物资储备的机构肩上。最终，监管链条全线失守，反倒是由官方媒体的调查报道率先揭开了黑幕。

中国正寄希望于科技手段来破解难题。有些措施简单直观：许多中国餐馆如今普及了“明厨亮灶”在线直播，食客可以实时查看菜品制作过程。另一些措施则更加前沿：部分大型企业利用区块链——一种安全共享的去中心化数据库——来对商品进行全流程溯源。中国还在大力推进“冷链”基础设施建设，即利用冷藏货车和温控仓库网络，使果蔬在运输途中无需依赖甲醛等化学防腐剂便能保鲜。政府近期的一项规划提出，在现有的1000个冷链物流中心基础上，未来五年内再增设500个。

然而，这些手段都会推高成本。富裕阶层似乎很乐意为此买单。美国零售巨头沃尔玛旗下的山姆会员商店，凭借提供安全优质食材的承诺在中国迅速扩张。阿里巴巴旗下主打本土创新的盒马鲜生，也允许消费者扫描部分商品二维码查看农残检测报告和全程运输日志。但并非所有人都能负担得起更加昂贵的生鲜账单。

按下葫芦浮起酱

要想让民众享受到既实惠又绝对安全的食品，中国必须解决更为深层的积弊。其一是企业文化。中国企业往往以产品推向市场的极高速度为傲，但这往往会导致对供应商的资质审核和其他安全风控流于形式，企业内部向管理层反映安全漏洞的渠道也普遍不健全。为此，政府正鼓励员工直接向监管部门举报。2024年，官方宣布提高食品行业“吹哨人”的奖励标准，在重大案件中最高可按罚没款金额的5%予以重奖。

但一位中国食品安全专家警告称，体制内的激励导向同样存在扭曲。基层官员若想获得提拔晋升，必须发展地方产业、保障本地就业。而对食品安全过于严苛的执法，可能会危及这两个目标。他指出：“大家都在食品安全与经济增长之间维持着一种心照不宣的平衡。”一旦爆出重大丑闻，通常会引发上级部门立案彻查，随后无可避免地揪出若干地方官员问责“背锅”。心知肚明的地方官员因此更倾向于内部消化问题，或者尽可能装作视而不见。

在人均GDP仅为北京三分之一左右的康保县，农业在该县经济格局中举足轻重。面对这场舆论海啸，当地官员想必正严阵以待，并且对新闻媒体保持着高度戒备。本刊记者抵达康保县后不久，一辆警车便悄然尾随而至，四名警员随即下车，紧接着又赶来一众宣传部门的官员。他们热情洋溢地大谈当地宜人的气候，以及在附近繁衍栖息的某种珍稀鸟类，唯独对记者的任何白菜问题三缄其口、避而不谈。■

ch017.xhtml

中国军事力量的远程辐射

中国战机为何在埃及上空举行联合空战演习

图片来源：Getty Images

2026年9月1日

8月最后一周，当六架中国战斗机飞越埃及时，它们所展示的既是一场精湛的空中特技，更是一场高调的外交战术。这些歼-16重型战斗机是中国空军航程最远、武器挂载能力最强的机型之一，此次是它们首次与埃及举行联合军事演习。考虑到埃及与美国长期保持着密切的防务合作，加之各方对美国在中东地区的可靠性产生新的疑虑，这次军演格外引人瞩目。更重要的是，中国战机首次与法制“阵风”战斗机展开模拟空战演练。埃及装备了该型战机，而与中国在2020年爆发致命边境冲突的印度同样大量列装了这款机型。

这场为期两周、代号为“文明之鹰”的联合演习预计将于9月初收官，届时中国最高领导人习近平将抵埃访问，这也是他时隔十年再度踏上埃及土地。演习与出访很可能在今年2月美伊爆发战争之前就已提上日程。这也是中埃连续第二年举行双边空中联合演习，而官方层面上，习近平此访旨在庆祝中埃建交70周年。然而，这一切依然向外界释放出了中国在该地区地缘雄心的明确信号。

中国正在展示其作为替代性军事装备供应国的巨大潜力，以及日益强大的远程空中力量投送能力。而中国展示这些能力的时机，恰逢美国在中东的军事主导地位受到严重动摇之际——美国未能击溃伊朗，自身武器库存严重见底，且未能有效保护屡遭打击的中东军事基地。

中国目前既无能力，似乎也无意愿取代美国在中东的霸权地位。但为了援助合作伙伴并保护自身在这一至关重要的贸易与能源枢纽的利益，中国在执行有限军事任务方面的能力正稳步提升。与此同时，中国借此获取了宝贵的一手经验和技战术数据，有助于其飞行员在未来应对“阵风”战机以及其他西方同代机型。这无疑会令印度高度警惕。在2025年印巴冲突期间，印度至少有一架“阵风”据信被巴基斯坦装备的中国制造战机击落（印度军方虽承认损失了数架战机，但否认其中包含“阵风”）。

埃及绝非中国在中东开展军事外交的唯一样板。中国还借助军售和联合军演，不断加强与沙特阿拉伯和阿拉伯联合酋长国等其他国家的防务合作，尽管这两个国家长期与美国保持着密切的军事纽带。中国与阿联酋原定于8月下旬在中国西北部启动新一轮空军联合演训，这是两军自2023年以来建立的年度机制。虽然双方均未披露参演的具体机型，但以往的演练曾涉及阿联酋装备的法制“幻影”战机，而中国台湾地区同样列装了该型战机。

这类联合演训在一定程度上兼具军贸推销的性质。据报道，埃及在2024年签署了首份中国战机采购订单，订购了数量未公开的单发歼-10战机（中国官方对此予以否认）。但在2025年中埃首轮联合空中演习中，中方派出了歼-10战机参演，显然有意借此锁定军售大单。埃及此前可能考虑用歼-10替代其自1980年起采购、现已日趋老化的美制F-16机队。如今，中国似乎正在大力推介航程更远、载荷更大的双发重型歼-16，以助力埃及在红海、非洲之角等周边空域执行更大纵深的战备巡逻任务。

此次演习也为中国检验其跨战区远程兵力投送能力提供了实战平台。歼-16机队飞行约9小时、跨越6000公里自中国直飞埃及，全程由一架随行的运油-20大型加油机提供空中加油保障。此外，中国战机及保障机群还在巴基斯坦卡拉奇进行了经停。巴基斯坦是中国最亲密的军事盟友之一，此次经停凸显出该国作为中国未来向更远区域投送军事力量的中转保障基地的巨大潜力。

不过，最引人注目的还是与埃及空军“阵风”战机的模拟空战。中国国家电视台公开了歼-16座舱视角拍摄的实况画面，清晰记录下伴随做对抗机动的一架“阵风”战机。官方报道称，演习涵盖了一对一和二对二的空战演练。相比同型机对抗，此类异型机对抗演练（DACT）能够为飞行员提供更具实战价值的空战经验，同时也让双方都能摸清对手战机的雷达电磁及飞行性能底细。中国究竟还要多久才能在远离本土的空域遂行高强度实战，目前尚无定论；但作为一个长远的战略目标，中国显然已经锁定了这一靶心。■

ch018.xhtml

严肃较劲的中国烧烤江湖

一些城市宣称，早在汉代自己就已经吃上烤串了

排队等串的长龙 © 图片来源：Getty Images

2026年9月3日 | 齐齐哈尔

德克萨斯烤肉得先往后稍稍了。中国城市齐齐哈尔正向全球烧烤之冠发起冲击。这座位于中国东北边陲的城市不久前刚落幕了一场烧烤美食节，向全球食客发出品尝风味的盛情邀请。在这座拥有400万人口的城市里，大街小巷随处悬挂着标榜自己为“国际烤肉之都”的横幅。对于一个在海外少有人知、更少有人能读对名字（大致读作 chee-chee-har）的地方而言，这无疑是一个极为宏大的头衔。

地图来源：《经济学人》

齐齐哈尔真正较劲的对手并不是德州牛胸肉或澳洲烤肠，而是中国国内的其他兄弟城市。近年来，烧烤已成长为一门体量巨大的朝阳产业。到2025年，中国烧烤餐饮市场整体规模已达约2700亿元人民币（约合400亿美元）。多座城市正使出浑身解数，试图在这一庞大市场中分得最大一杯羹。在这场角逐中，齐齐哈尔底气十足。这里拥有中国最辽阔的优质牧场之一。每逢盛夏，当地居民便三五成群围坐街头，将肉香四溢的拌肉放在炭火烤盘上，伴着洋葱和酸菜滋滋作响。齐齐哈尔人坚信自家的牛肉天下第一，并且拿得出实打实的数据：从2020年到2025年，该市的牛肉产值已从约50亿元狂飙至300亿元。

对齐齐哈尔构成最大冲击的挑战者来自淄博，一座位于中国东部、以小饼卷烤串闻名的城市。2023年，淄博烧烤摊一跃成为全国现象级旅游目的地，当地旅游收入同比暴增70%。地方官员迅速顺水推舟跟进扶持，鼓励银行发放“烧烤贷”，帮助创业者采购羊肉、支起摊位；由于客房一房难求，政府甚至紧急出台了酒店限价令。

淄博的爆红引爆了一场更大规模的全国“烧烤大战”。另一座东北城市锦州多年来一直自诩为中国的小串发源地；看到淄博的造富神话后，锦州官员火速派出考察团登门“取经”，试图把自家的烧烤招牌变现。而东部另一座重镇徐州的市民则认为，自家的烤肉渊源远比淄博悠久，理应获得更多游客青睐。当地出土的两千年前汉代画像石上，赫然刻有先民炙烤肉串的生动图案；当地网民甚至愤愤不平地抨击淄博窃取了徐州的非遗文脉。

中国夏季的网红旅游地往往在社交媒体上各领风骚三五月，随后热度迅速冷却。淄博如今已难掩风潮退去后的落寞，据报道，当地对烧烤行业的大干快上导致不少后入局的商户折戟沉沙。相比之下，齐齐哈尔如今依托遍布全国的加盟连锁餐饮网络销售本地牛肉，从而为地方实体经济注入了持久动能。齐齐哈尔或许暂时无法撼动德克萨斯烟熏坑烤老手的地位，但在问鼎中国烧烤之都的赛道上，它绝对握有极具分量的筹码。■

ch019.xhtml

银幕与现实：中国推行“龙餐厅外交”

虽充满矛盾，却被许多国家视为抗衡西方霸权的另一种解法

插画：Cornelia Li

2026年8月31日

以虚构伊拉克为背景的中国最新票房大片《欢迎来到龙餐厅》（《莫斯科行动》团队打造，英文片名直译为《从前在中东》），大概率与奥斯卡奖无缘。该片剧情荒诞夸张，不少角色脸谱化严重。然而，海外观众却很有必要看一看这部电影。原因之一在于，它从一个外界鲜少接触的独特视角，重现了2003年美军入侵伊拉克的惨烈历史——那绝非西方家庭电视荧屏上泛着绿色夜视滤镜的高科技光影秀，而是重磅航弹砸向沉睡都市时撕心裂肺的呼啸与轰鸣，真实而残忍。

影片跟随一位在战火中发国难财、最终却被战火反噬的中国厨师的轨迹，展现出耐人寻味的政治寓意。它折射出中国如何看待自身形象，以及其在中东乃至全球舞台上的战略定位。如果说对一部商业电影进行如此过度解读显得有些小题大做，不妨回想一下，过去十年间外界曾普遍将另一部电影《战狼》视为中国咄咄逼人的进攻型外交姿态的代名词。相比之下，“龙餐厅外交”则折射出一种截然不同的风格——它甚至能更准确地勾勒出当今中国外交的全貌。其核心运作准则在于广结善缘、不树私敌，其底层驱动力与其说是政治意识形态，不如说是务实的重商主义。这种行事风格无论在电影里还是现实中，往往浮于表面且充满自相矛盾；但对于那些苦西方傲慢久矣的发展中国家而言，低估其广泛的吸引力将是大错特错。

影片主角名为徐福。他远走虚构的巴格达，原本只是为了靠厨艺多赚点钱。战争爆发后，他敏锐地嗅到了驻伊美军驻扎的“绿区”内蕴藏的巨大商机。西方大兵以及——咳咳——战地记者们，对他的中餐赞不绝口，对他的美酒更是趋之若鹜。徐福赚得盆满钵满，直到一场自杀式爆炸袭击彻底打破了平静。他的前合伙人牵涉其中，导致徐福沦为美军阶下囚；获释后不久，他又被极端武装分子绑架。中国官媒和知识界将该片盛赞为一则政治寓言，意在昭示饱受美国摧残的世界能从中国那里获得什么。然而，这部电影蕴含的意涵远比这复杂得多，其中有些切面甚至并不完全是在为中国唱赞歌。

影片的核心立意在于：中国希望博得所有人的好感。徐福堪称一名八面玲珑的中国职业外交官的化身：他的当地合伙人是战前萨达姆政权的体制内官员，即便政权崩塌、对方失势，徐福依然对其不离不弃；与此同时，美国官兵也逐渐接纳了他，尤其痴迷于他的手艺；就连劫持他的恐怖分子头目，也最终与他相谈甚欢。面对所有人，徐福看到的都是共通的人性。值得肯定的是，电影并未回避这种圆滑背后的深层张力。一名美国士兵曾当面质问他：“你的合伙人是加害者，你的朋友是受害者。你真觉得自己跟这场战争毫无干系吗？”这番诘问，或许也足以引发那些在俄乌冲突中坚称保持中立、却在面对西方指责中方经贸往来支撑了俄军战力时恼羞成怒的中国官员的深思：你真觉得自己能独善其身吗？

徐福的行事动机同样十分真实传神。从一开始，为了还债而渴望捞快钱的他就坦率得甚至有些不加掩饰。正如他的中国合伙人所感慨的那样，能在战争中发财纯属老天赏饭吃。这是驱动中国海外商业活动的那种机会主义实用倾向的一个生动缩影。这听上去或许不够崇高，但对于身处接收端的对象国来说，这种姿态往往大受欢迎——它与人们普遍认为西方资金总是附带苛刻政治条件的印象形成了鲜明反差。中国科技巨头华为创始人任正非的名言“在商言商”，正是这一逻辑的精辟概括。摒弃政治意识形态的干扰，正是中国能够在错综复杂的中东乃至更广阔世界中左右逢源、跨越敌对阵营鸿沟的关键密码。试看伊朗与海湾君主国、阿尔及利亚与摩洛哥、埃及与埃塞俄比亚——面对这些地缘死结，绝大多数国家都不得不选边站队，中国却能超然其外。在许多国家看来，与中国打交道的人都乐于接受这一套说辞，哪怕“政治为商业让路”有时只是一种美好的幻象。

然而，“龙餐厅外交”的另一个鲜明特征在于其内核的单薄。在影片后半段，徐福被赋予了拯救战乱孤儿的人道主义光环。但他自始至终都是一个极不情愿的救世主。片中一位朋友曾劝告他：“你本该置身事外的，他们又不是你的骨肉。”最终，他听从了这番劝诫，收拾行囊回到中国，将战乱废墟彻底抛在身后。用徐福自己的话来说，电影的终极反战主旨不过是“好好吃饭，好好把孩子养大”。这话本身挑不出毛病，但它恰恰折射出中国对海外伙伴支持力度的局限性。中国对卷入外部纷争——尤其是军事层面的泥潭——抱有极高的警惕。伊朗和委内瑞拉在今年对此深有体会。“这是一个管理预期的问题，”中阿关系专家卡里姆·瓦迪（Karim Alwadi）指出，“中国根本无意扮演别国的安全担保人。”

大厨的招牌菜

对于那些曾对中国“战狼”式强硬转身感到寝食难安的西方人而言，“龙餐厅外交”的兴起似乎是一个巨大的改观。它战线拉得极长，看似人畜无害，甚至略带天真。但实事求是地讲，这是一种更为严峻的制度性挑战。不管存在多少内在矛盾，对于广大发展中国家而言，这种模式也远比动辄诉诸关税制裁、甚至武力轰炸迫使别国屈服的西方手段更具亲和力。毕竟，只要真金白银的投资能够落地，就永远比高高在上的道德说教更受欢迎。

然而，中国自身也有被这套自制迷汤麻醉的风险。该片在现实中的原型之一是刘磊，他于2003年辞去IT工作前往巴格达开餐馆。在他合著的回忆录《在死神脚下赚钱》中，真正体现崇高人道主义光辉的不是刘磊本人，而是一名年轻的日本反战志愿者。这名日本女孩在他店里当服务员，每天下午都会把客人的剩菜收集起来送往当地孤儿院。然而，这样的剧情线显然很难通过中国严格的影视审查。在试图与全世界所有人交朋友的宏大叙事中，中国唯独留下了一个刺眼的例外——日本。中国领导人频繁指责日本军国主义死灰复燃，似乎绝难承认日本人对和平的渴望可能丝毫不亚于中国所标榜的诚意。在这场“龙餐厅外交”盛宴中，究竟该上什么菜，解释权永远牢牢掌握在中国自己手中。■

ch020.xhtml

数据中心何以演变为全美最棘手的政治焦点

民间抗议浪潮愈演愈烈，正在重塑中期选举版图

图片来源：Getty Images

2026年9月2日 | 芝加哥与华盛顿特区

一个幽灵，一个数据中心的幽灵，正在美国上空徘徊。这些摆满服务器的钢筋水泥仓库是驱动美国人工智能算力热潮的心脏，如今却沦为公众接受度最低的基础设施——其民意支持率甚至跌破了核电站。全美各地的县市和州政府正竞相用繁文缛节对其进行层层设限，乃至全面下达封杀令。今年7月，旨在限制数据中心推高民用电价的《受规管用电户保护法案》（Ratepayer Protection Act）在联邦众议院委员会层面一路绿灯顺利通过。

8月31日，唐纳德·特朗普介入了这场白热化的争议，令深陷中期选举胶着选情的共和党人倍感焦虑。他发出严厉警告称，如果地方阻挠势力杀死了这只“下金蛋的鹅”，他们最终将陷入“闭塞与贫困的泥潭”。民间针对数据中心的强烈抵触情绪，正在撕裂传统的党派政治同盟，重塑即将到来的选举格局，也令科技巨头和投资机构坐立难安。至于这场反弹风暴是否会遏制势不可挡的人工智能发展列车，目前尚不得而知。

要观察数据中心的政治风向转变有多迅猛，伊利诺伊、德克萨斯和宾夕法尼亚这三个政治底色截然不同的州便是绝佳样本。2019年，伊利诺伊州州长J·B·普里茨克（J.B. Pritzker）签署了一项两党支持的法案，为数据中心建设提供税收优惠。他当时宣称：“在当今世界，数据中心与我们的公路、铁路和学校一样，都是不可或缺的基础设施。”在德克萨斯州，州长格雷格·阿博特（Greg Abbott）直到去年还在热烈欢迎谷歌新建数据中心，并自豪宣称德州是“全球AI发展的核心枢纽”。宾夕法尼亚州州长乔希·夏皮罗（Josh Shapiro）此前同样热情高涨：“我们已经全方位拥抱AI。”

然而到了今年8月，普里茨克州长的话锋突变。他向媒体直言，数据中心带来了“推高电费、水费或破坏生态环境”的巨大隐患。早在6月，他就已单方面叫停了2019年法案所确立的税收抵免优惠申请。在德克萨斯州，阿博特州长下令州电力监管机构对所有申请接入电网的数据中心开展全面审计。他警告称，这些设施“可能严重危及德州电网的可靠性与稳定性”。夏皮罗州长同样收紧了政策门槛，赋予地方社区更大的项目一票否决权，同时在州长竞选中给对手贴上“亲数据中心”的标签——在当下，这已日益沦为一种政治污名。

图表来源：《经济学人》

就在一年多以前，部分民调仍显示过半数民众支持在本地兴建数据中心。但到了今年3月，民意发生彻底逆转：盖洛普（Gallup）的一项民调显示，反对比例已高达71%。安嫩伯格公共政策中心（Annenberg）的追踪调查显示，从2月到6、7月间，数据中心的好感度净值狂跌了19个百分点。《经济学人》与YouGov在8月底联合开展的民意调查显示，反对率攀升至63%。在几乎所有选民群体眼中，数据中心都被普遍视作国家发展的祸患（见图表1）。

芝加哥的剧变极具代表性。得益于地处美国互联网骨干网络核心枢纽的区位优势以及相对廉价的电力供应，早在AI浪潮席卷之前，芝加哥市区及其周边郊县就已聚集了海量数据中心。长期以来，当地民众对此鲜有异议。但如今，抗议者挤满了大大小小的市政听证会，力图叫停全州境内的所有开发项目。8月，数百名群众自发聚集抗议芝加哥南部的一处规划方案——令人啼笑皆非的是，据开发商透露，该项目根本就不是数据中心。

图表来源：《经济学人》

今年前八个月，全美各县审议的暂缓营建令或类似限制性措施高达409项，几乎是去年全年的七倍。在六个以上的州——其中包括几乎所有的中期选举摇摆战场——超过五分之一的县已表决通过了至少一项暂缓营建令（见图表2）。达拉斯温斯特德律师事务所（Winstead）土地规划律师汤米·曼（Tommy Mann）感叹道：“从土地使用律师的执业生涯来看，这是我经手过的阻力最大的开发纠纷之一。围绕这些项目所爆发的情绪对抗激烈程度，我前所未见。”

部分数据中心倡导者将民意的反弹归咎于境外势力的蓄意挑拨。特朗普评论称：“中国乐见其成。”今年6月，人工智能独角兽OpenAI宣称，一家中国公司炮制了一批煽动性图片，旨在瓦解美国本土对数据中心建设的民间支持。8月27日，社交媒体平台X声称捣毁了一个由20万个虚假账号组成、负责散布上述素材的中国“僵尸网络水军”。然而，既有事实表明这些境外策动收效甚微。OpenAI最终得出的结论是，这些内容在普通美国网民中“几乎没有引发任何可观测的实际互动”。

邻避效应的蔓延

民怨沸腾的症结实则要现实得多。美国人普遍对人工智能心存疑虑，缺乏对科技巨头的基本信任，并对算法取代就业深感恐惧。但多项民调一致表明，让基层选民夜不能寐的，核心在于数据中心对社区周边的现实冲击——或者说，在于他们脑海中所假想的破坏。福克斯新闻（Fox News）近期的一项民调揭示，高达75%的受访者将能耗激增、水费上涨和交通拥堵列为反对本地建设数据中心的首要理据，仅有11%的人将矛头直接指向AI技术本身。

许多担忧在很大程度上被无限放大了。超大规模云服务商谷歌指出，2025年其全球数据中心和办公区消耗了约109亿加仑（约合410亿升）的水资源——据其测算，这大致相当于美西南部73个高尔夫球场一年的灌溉用水。另一家巨头亚马逊表示，尽管去年旗下的数据中心集群规模进一步扩张，但其全球水耗总量为25亿加仑，同比反而缩减了2%。纵观2023年，即便算上为数据中心发电所耗费的水量，全美农田灌溉用水量依然是数据中心的74倍。

然而，用电负荷则是完全另一码事。联邦资助的劳伦斯伯克利国家实验室测算，2023年数据中心用电量已占全美总电力的4.4%，预计到2030年这一占比将飙升至9.5%到15.3%之间。其对终端消费者的传导效应尚存争议。批发端电价上涨确实已成既成事实，并终将向下游家庭电费账单传导；但美国电力科学研究院（EPRI）的研究指出，从2019年到2024年，数据中心的爆发式增长通过大规模分摊电网固定建设成本，反而使全美平均电价下降了6%。

最近几周，佐治亚电力（Georgia Power）、太平洋瓦电（PG&E）以及印第安纳密歇根电力（Indiana Michigan Power）三大电力公用事业巨头均公开表态，坚称数据中心的接入有助于平抑居民电费。与此同时，已有30个州批准或提议征收“大负荷用电附加费”，以确保数据中心等耗电大户切实自掏腰包为电网升级改造买单。

核心悬念在于民意是否仍有回旋余地。民调机构Echelon Insights本月公布的调查显示，只要告知受访者其中哪怕一项切实利益——例如降低个人财产税或加大对公立学校的资金倾斜，公众对数据中心的反对立场就会迅速瓦解。Echelon负责人、共和党民调专家帕特里克·鲁菲尼（Patrick Ruffini）认为：“选民完全愿意就数据中心达成利益妥协。”不仅如此，经济补偿同样立竿见影。弗吉尼亚大学开展的另一项民调发现，大多数选民只要拿到几千美元的补贴就愿意接纳本地数据中心，要价的中位数大约为1万美元。按全美中等规模县城的人口基数计算，给每位居民发钱的开支约为2.4亿美元——对于财大气粗的硅谷科技圈而言，这不过是九牛一毛。事实上，巨头们如今早已开始向当地社区密集撒钱。

尽管如此，抵制浪潮短期内依然难言退潮。布鲁金斯学会智库学者达雷尔·韦斯特（Darrell West）指出：“到2028年，这番攻防大戏还会卷土重来。”他分析称，今年侥幸躲过该议题敲打的州长们，在两年后的连任大选中注定避无可避。

数据中心的政治博弈在其他维度同样荒诞颠倒。对于工会势力根深蒂固的伊利诺伊州民主党人而言，民意的反噬正让他们陷入进退维谷的绝境。许多工会组织是数据中心的坚定拥趸。在普里茨克州长6月做出叫停决定后，数家工会领袖联名发表谴责声明，痛批此举“无异于将数十亿美元的产业投资和数以千计的工会岗位拱手让给印第安纳、肯塔基和俄亥俄等邻州”。当地工会官员蒂姆·吉鲁利（Tim Gillooley）直言不讳：“这些数据中心早晚都要落地！为什么不建在我们家门口？”

阻力最小的阻击路径

与日俱增的阻力宛如给狂飙突进的产业列车掺进了泥沙。半导体芯片作为新建数据中心中最大的一笔资本支出，其销售税免税政策在德克萨斯等绝大多数州依然坚挺。但在亚利桑那和俄亥俄等州，相关免税已被紧急叫停，宾夕法尼亚州则为其设立了繁复的先决条件。这一趋势叠加不断攀升的本地安抚成本，必将推高新建数据中心的投资造价。即便如此，少数几个核心州依然握有绝对的定盘话语权。仅德克萨斯一州在建的数据中心体量，就超过了佐治亚、亚利桑那和宾夕法尼亚三州的总和（见图表3）。

图表来源：《经济学人》

资本最终可能会顺着阻力最小的方向，流向监管最为宽松的庇护所。咨询机构Capstone能源分析师杰克·佩因特（Jack Painter）总结道：“只要能带来可观的财政税收和其他衍生红利，总会有足够多的地方社区张开双臂欢迎这些项目落户。”正如特朗普所言：“让数据主宰一切。”■

ch021.xhtml

美国政治怒火的量化剖析

国家正深陷集体愤懑的泥潭

图片来源：Getty Images

2026年8月30日

仅在过去两个月里，唐纳德·特朗普就频频将民主党候选人贬斥为“圣战分子”和“激进疯子”。他肆意嘲弄政治对手的体型与长相，公然宣称自己痛恨来自加利福尼亚州的民主党参议员亚当·希夫（Adam Schiff），还将德克萨斯州民主党参议员候选人詹姆斯·塔拉里科（James Talarico）蔑称为“塔拉怪胎”（Tal-a-freak-o）。不仅如此，他还对昔日的共和党同僚反戈相向，斥之为“彻头彻尾的失败者”。

尽管他的言辞攻击性远甚于绝大多数政客，但这位前总统绝非唯一贩卖愤怒的政客。最新学术研究表明，他只是庞大历史洪流中的一个缩影：当今美国政治正处于前所未有的暴怒周期。

来自哈佛大学、巴黎高等商学院（HEC Paris）、巴黎萨克雷大学以及加州大学伯克利分校的经济学家们，联手展开了一项关于政治怒火演变轨迹的前沿研究。他们梳理了150多年来的美国国会演讲实录以及数以百万计的社交媒体推文，利用大语言模型和自然语言处理等人工智能工具，对文本的情感属性进行了系统性分类标定。研究团队对每段语料在愤怒、悲伤、自豪、感激、恐惧、厌恶、喜悦和希望这八种核心情绪维度上逐一打分，并测算其整体情绪化烈度。

图表来源：《经济学人》

研究报告揭示，国会辩论中的愤怒言论在20世纪90年代纽特·金里奇（Newt Gingrich）掌舵众议院期间呈爆发式增长，并在本世纪初短暂回落。然而近年来，这类言论卷土重来且愈演愈烈，自2013年以来暴增了50%（见图表）。选民的戾气同样水涨船高。从2013年到2025年，普通选民在社交平台推文中涉及公共政策且带有愤怒情绪的语句占比攀升了35%；在网络论坛Reddit的跟帖讨论中，愤怒言论的比例更是猛增了46%。

该论文的作者们在归因时十分审慎，并未将怒火的蔓延单一归咎于某种特定诱因，尽管学界对此见仁见智。印第安纳大学的史蒂文·韦伯斯特（Steven Webster）强调了党派对立极化的深远影响。皮尤研究中心（Pew Research Center）的统计显示，自2016年以来，两党选民不仅在政见上水火不容，更在道德层面将对方阵营妖魔化为不道德与不诚实的化身。从2016年到2025年，公开宣称对方党派让自己感到“愤怒”的铁杆党民比例激增了近30个百分点，达到惊人的76%。

该论文联席作者、哈佛大学的斯特凡妮·斯坦切娃（Stefanie Stantcheva）指出：“无论这一循环始于何处，它最终都会沦为一个自我强化的恶性反馈闭环。”她援引数据证明，国会议员发布的充斥愤怒情绪的推文，其转发量往往比情感中立的推文高出59%。与其他情绪相比，愤怒更能高效动员选民走出家门投票，进而启动一种极具破坏力的恶性循环。政治学者早已证实，即使怒火的源头与政治本身毫不相干，群体性狂怒也会持续侵蚀公众对政府的信任，加剧阵营极化撕裂，并削弱选民对民主制度规范的认同坚守。

斯坦切娃博士及其团队进一步发现，当受访者被诱发负面情绪后，他们对移民和自由贸易的包容度显著下滑，而对财富再分配政策的支持率大幅上升——这或许是社会底层仇富心态的一种投射。她解释道：“当你处于愤怒状态时，你大脑下意识调取的认知框架就只剩下委屈怨恨与追责归咎。”

随着中期选举的临近，在政客竞相争夺眼球的厮杀中，激化愤怒已成为许多竞选活动的核心底牌。今年4月，新闻网站Axios报道称，近50支民主党竞选广告中赫然充斥着“斗士”（fighter）或“战斗”（fight）等字眼。在科罗拉多州，独立参议员候选人鲍勃·丘（Bob Chew）甚至投放了一则自己在房间内掀桌砸椅的竞选广告，借此宣泄对生活成本高企的滔天怒火。

亦有少数候选人试图另辟蹊径。民主党参议员候选人乔希·图雷克（Josh Turek）表示自己将“为爱荷华而战”，但他同时极力呼吁“重拾希望与正能量”。塔拉里科则公开倡导以公仆之心战胜自私自利。去年，佐兰·马姆达尼（Zohran Mamdani）正是凭借民粹主义政纲与永远洋溢着乐观笑脸的亲民形象，一路过关斩将爆冷问鼎纽约市长宝座。美国选民的心头依然积郁着厚重的怨气——呼唤彻底变革的民意浪潮空前高涨。然而，并非每位政客都坚信，紧锁双眉、咬牙切齿才是指引前路的最佳姿态。■

ch022.xhtml

皮特·海格塞斯正在掌控五角大楼

其最新一轮清洗终结了陆军改革领军人物的仕途

图片来源：Getty Images

2026年9月1日 | 华盛顿特区

一年半以前，几乎没有人看好皮特·海格塞斯（Pete Hegseth）能在五角大楼彻底大权独揽。此前，这位前福克斯新闻台主持人兼高调的“文化战争”旗手曾深陷过往任职期间财务违规以及酗酒等丑闻指控（他对此予以坚决否认），并仅仅以微弱优势惊险通过参议院的提名确认。在外界看来，他履历浅薄、资历平平，根本无力节制麾下那些身经百战的将军和舰队司令。然而走马上任以来，这位战争部长虽频频游走于丑闻边缘——从涉嫌泄密到强令下属接受测谎——但其权势不仅没有被架空削弱，反倒将国防部牢牢掌控在自己手心。通过在全军掀起一轮接一轮的清洗风暴，他已将潜在的政敌和被视作忠诚度欠佳的高级将领彻底清扫出局。

陆军部长丹·德里斯科尔（Dan Driscoll）成为这轮清洗的最新牺牲品。8月31日，在就清洗高级将领以及陆军未来建军方向与海格塞斯爆发长达数月的激烈对抗后，德里斯科尔正式向总统递交辞呈。作为一名备受军内外爱戴、被公认为极具远见卓识的改革派领军人物，德里斯科尔的黯然挂冠，让五角大楼内部的权力震荡愈演愈烈，也给旨在推动陆军加速转型以适应无人机战争时代的各项重大军事改革蒙上了一层厚厚的阴影。

海格塞斯的整肃手段可谓疾风骤雨。自履新以来，他已罢免了至少24名将官。许多人遭到免职的缘由，似乎仅仅是其种族、性别，或是因其与海格塞斯看不顺眼的将领走得过近。作为一名曾官至国民警卫队少校的退伍军人，海格塞斯长期对陆军建制派怀恨在心，他曾公开控诉这个体制当年“将我无情抛弃”。今年4月，他悍然解职了陆军最高长官、陆军参谋长兰迪·乔治（Randy George）上将，而后者正是德里斯科尔生前最紧密的改革盟友。仅仅两个月后，在海格塞斯强行降低美国驻欧非陆军司令部规格后，其司令克里斯·多纳霍（Chris Donahue）上将被迫申请提前退役。时至今日，美国陆军已陷入既无参议院正式确认的参谋长、亦无正式陆军部长、更无正式欧洲部队司令的空前瘫痪状态。据传德里斯科尔的副手也已萌生去意、离职在即。

德里斯科尔曾竭尽所能筑起阻击防线。一名国会资深助理透露：“他此前成功保下了多位险些遭到解职或晋升受阻的将领。”然而，海格塞斯最近强行从18名拟晋升陆军少将的名单中划掉了7名候选人，这或许成为了压垮骆驼的最后一根稻草。

这轮人事地震爆发的时机极为敏感。作为副总统J·D·万斯（J.D. Vance）的密友，德里斯科尔长期被视为抗衡并制约海格塞斯激进鹰派倾向的关键刹车阀。就在递交辞呈的前几天，他还曾当面向唐纳德·特朗普当面陈情，集中反映了五角大楼内部因应对美伊战争而导致军队战备高度紧绷、以及对海格塞斯插手军队人事任免的普遍焦虑。

急剧的人事动荡也严重阻碍了德里斯科尔生前牵头的多项里程碑式现代化改革。他一直是五角大楼无人机装备采办体系的核心掌舵人（特朗普曾调侃地称他为“无人机小子”）。去年11月，他亲赴乌克兰，与当地顶尖无人机专家建立了技术联合攻关机制。然而，德里斯科尔积累的巨大声望据信引发了海格塞斯的强烈嫉恨，后者正急于将其及其盟友推行的一系列先锋试点项目全盘推倒重来。例如，乔治上将此前为了摸索现代无人机战术战法而专门设立的一个全新陆军数字化试验营，目前正被深得海格塞斯器重的代理陆军参谋长克里斯托弗·拉内夫（Christopher LaNeve）中将逐步关停裁撤。

拉内夫中将更热衷于高谈“重归基础训练”。尽管其口头上并未放弃采购前沿科技装备，但资源重心已明显向“让美国再次伟大”（MAGA）式的政治诉求倾斜。拉内夫不仅主导了美军从欧洲战区的撤军收缩，还将更多兵力部署转向西半球。此前，德里斯科尔还就推倒重写全军体能测试标准一事与拉内夫彻底闹翻，而体能考核改革恰恰是海格塞斯多年来念念不忘的心头大患。

这位国防部长极有可能进一步巩固其绝对权威。五角大楼现任发言人肖恩·帕内尔（Sean Parnell）被普遍视作接替德里斯科尔的热门人选。与其上司如出一辙，帕内尔同样痴迷于投身文化战争。尽管其匮乏大型组织管理经验的短板可能会令参议院的任命确认困难重重，但他完全可以长期以代理部长的身份发号施令。随着时间一天天推移，整座五角大楼正被彻头彻尾地雕琢成海格塞斯个人的政治翻版。■

ch023.xhtml

加拿大反制美国的杀手锏在何处

其关税反击经过精确计算，旨在释放最大的政治威慑

身处风暴前沿 © 图片来源：Getty Images

2026年9月3日 | 华盛顿特区

唐纳德·特朗普对加拿大挑起贸易战的底层逻辑极其简单粗暴：美国手握所有王牌。“我们根本不需要加拿大，是他们离不开我们，”特朗普于8月24日在社交媒体上直言不讳地写道。一周后，美国财政部长斯科特·贝森特（Scott Bessent）在面对加拿大的关税报复威胁时同样不屑一顾：“面对一个经济体量是你13倍的对手，你根本打不起针锋相对的贸易战。”

然而，这一论调严重低估了美国对其最核心商业伙伴之一的依赖风险。两国去年的双边货物与服务贸易总额突破了8700亿美元。加拿大消化了美国约15%的商品出口，更是全美34个州的第一大海外出口市场。在汽车制造领域，零部件在最终组装成整车之前，往往需要在美加边境往返穿梭多次。从美国宏观基本盘来看，这场贸易战的波澜似乎微不足道；但对于特定关键行业和部分关键州而言，代价却堪称致命。

加拿大正处心积虑地将这些经贸纽带转化为高精准度的政治杠杆。面对美方对价值200亿美元的加拿大出口产品征收50%关税的做法，渥太华方面给出的反制方案是对等额200亿美元的美国进口商品加征15%至50%不等的报复性关税，定于9月8日正式生效。钢铁与铝制品赫然列入顶格征税名单，奶酪和农用机械等高敏感产品同样在劫难逃，汽车产业也将继续承担加拿大原有的反制性关税。“我们在精准挑选能够重创特定关键州的针对性产品，”加拿大工业部长梅拉妮·若利（Mélanie Joly）于8月25日毫不掩饰地挑明了意图。

加拿大的定点打击策略在中西部铁锈带体现得淋漓尽致。密歇根、俄亥俄和爱荷华三州眼下均面临胶着的中期选举参议院席位争夺战——并且无一例外都将结结实实地感受到这记重锤。针对钢铝产品新课征的50%惩罚性关税将直接重创密歇根州的装备制造业，而波及面更广的贸易争端则令当地汽车支柱产业雪上加霜。在爱荷华州，对农机设备征税将给约翰迪尔（John Deere）等本土农机制造巨头带来沉重打击。与此同时，俄亥俄州近三分之一的商品出口直接销往加拿大，极易首当其冲沦为炮灰。

尽管加拿大将海产品从反制清单中剔除，使缅因州的龙虾养殖业逃过一劫，但该州整体经济命脉与北方邻国依旧唇齿相依。木材、纸浆等初级林业产品多年来在两国边境之间往来流转加工。面对连任苦战的缅因州共和党参议员苏珊·柯林斯（Susan Collins）直言不讳地痛批白宫的关税政策是“一个巨大的历史性错误”。向南看去，对美国家具产品加征的50%关税，将精准重创北卡罗来纳州的家具产业集群，而该州同样有一场悬而未决的参议院公开席位争夺战正处于白热化关头。

加拿大的报复并非共和党在中期选举中面临的唯一梦魇。特朗普单方面开征的高额关税本身就将急剧推高美国国内的通胀物价。此轮关税覆盖了从儿童玩具到基础金属工业原料的方方面面，既推高了终端消费品的进口开销，也抬升了美国本土制造的生产成本。在生活负担成为全美选民第一核心关切的当下，这一自伤性政策无疑是在玩政治自焚。民调显示，全美仅有26%的选民支持特朗普对加拿大加征关税。

部分共和党建制派已开始公开拉响警报。8月24日，即将退休的北卡罗来纳州共和党参议员汤姆·蒂利斯（Thom Tillis）警告称，这场贸易龃龉正在边境各州引发极大的“民怨撕裂”。他意味深长地指出：“在11月的大选中，谁挑起了这种撕裂，谁就注定要吞下被选民清算的苦果。”其他议员则坚决反对特朗普将加拿大妖魔化为经济对手的做法。两天后，柯林斯参议员公开反驳道：“这可不是什么中国，这是加拿大，是我们最亲密无间的朋友，是一个与我们经济骨肉相连的同盟国。”

博弈成本甚至可能进一步升级。美国贸易代表杰米森·格里尔（Jamieson Greer）已公开扬言要对部分加拿大关键商品祭出全面进口禁令。而在加拿大国内，针对输美原油和关键稀缺矿产实行出口配额限制等极端反制手段，也已被摆上了决策层的研讨桌面。加拿大总理马克·卡尼（Mark Carney）态度坚决，重申渥太华只是被迫对美方的无端霸凌做出必要反击。贝森特财长或许说得没错，美国确实坐拥雄厚得多的关税弹药库，但这绝不意味着美国自身能够毫发无伤地从这场自残式战火中全身而退。■

ch024.xhtml

从身份转向阶级：觉醒文化2.0的蜕变

这无疑是一种进步

插画：David Simonds

2026年9月3日

承诺要为身处困境的纽约市民带来“集体主义温情”的纽约新任市长佐兰·马姆达尼（Zohran Mamdani），究竟能否开启一个均贫富的丰裕新纪元？目前尚无定论。然而，他麾下年轻的新一届市政班底，却早已给一个特定受众群体奉上了一份天降大礼——那就是该市保守派小报《纽约邮报》的拥趸们。这家小报正乐此不疲地大肆曝光马姆达尼在美国民主社会主义者（DSA）党内一众亲密盟友的奢靡生活，以此向右翼读者不断印证他们对激进左翼的鄙夷偏见。

“反富豪旗手、DSA领袖古斯塔沃·戈尔迪略（Gustavo Gordillo）寄生在百万富翁父母全款购置的150万美元曼哈顿豪宅中穷奢极欲，”《纽约邮报》上个月大肆喧嚣道。在极力嘲弄DSA“整天歇斯底里声讨私有产权”的同时，该报绘声绘色地大起底那栋由“其爹妈通过皮包公司代持”的奢华联排别墅中“修剪极尽考究的花园绿化”。《纽约邮报》随后又穷追猛打，接连曝出这位DSA纽约分会联席主席曾在耶鲁大学主修艺术与文学；曾因无故旷工被电工工会学徒项目扫地出门；其母亲平日更是以一辆银色保时捷代步。

沦为《纽约邮报》政治靶子的左翼青年远不止他一人。高呼“彻底废除房租”的特雷西·罗森塔尔（Tracy Rosenthal）同样被扒出家底——其父亲是一位功成名就的知名音乐制作人（报道标题直接赫然写道：“霸占布鲁克林房产的废除租金激进分子、马姆达尼闺蜜，实为格莱美得主的裙带巨婴”）。该报还将“来自俄亥俄州的25岁巧克力家族女继承人”格蕾丝·赖恩（Grace Ryan）痛批为DSA党内“最厚颜无耻的伪君子之一”，起因是她在社交平台上大秀自己身穿价值400美元的拉夫劳伦星条旗毛衣在私人游艇上吞云吐雾抽雪茄的自拍照。面对《纽约邮报》毫不掩饰的幸灾乐祸，曾在一场国会初选中为某位左翼大获全胜的候选人效力的赖恩女士，在社交平台X上反唇相讥：“怎么，难道你们这群人就从来没听说过‘富人当尽之责’（noblesse oblige）吗？”

这早已是一出上演了上百年的老剧码。自从卡尔·马克思不仅在共产主义理论上依仗弗里德里希·恩格斯，更在现实生计中依赖这位挚友丰厚的纺织厂家族资本供养以来，左派寄生于大资本家、右派据此极尽冷嘲热讽，就成了某种历史常态。自诞生之初，社会主义对某些沉迷标新立异的资产阶级时髦客而言，始终散发着一种带有叛逆色彩的别样魅力。然而，更为严谨清醒的社会主义学者则一针见血地将其视作该思想流派无法痊愈的阿喀琉斯之踵。乔治·奥威尔（George Orwell）在《通往威根码头之路》中曾辛辣地抱怨道：“社会主义最糟糕的名片，恰恰就是它自身的狂热信徒。”在奥威尔笔下，典型的社会主义拥趸要么是“年轻稚嫩、自命不凡的布尔什维克纨绔”，要么是“顶着白领光环、举止拘谨小家子气的小职员”，这些人大多“嘴上挂着素食主义高调”，最要命的是，“心里对自己既得的优渥社会阶层地位根本舍不得放弃半分”。

然而，赖恩女士关于“富人当尽之责”的辩解，倒真有一语中的之处。她或许一辈子也无法真正体验底层平民的艰辛，但这凭什么剥夺她试图向劳苦大众伸出援手的资格？这难道不比另一种典型形象强得多——一个身穿400美元名牌毛衣、对底层疾苦冷酷漠视甚至一心盘剥下层人的富家千金，究竟哪一种更具社会危害？

剥开这层政治作秀的浮沫，人们会触及一个至关重要的现实命题：左翼社会运动正在清晰地分化为界限鲜明的两个代际阶段，即外界所称的“觉醒文化1.0”（Woke 1）与“觉醒文化2.0”（Woke 2）。这种划分注定是模糊且具有过渡性质的，毕竟极少有民主党高层政客有足够的担当公开发表声明，同自己昔日深度绑定的极端立场决裂。一些人对过去的激进过火轻描淡写一笑了之——例如国会众议员亚历山德里娅·奥卡西奥-科尔特斯（Alexandria Ocasio-Cortez）引用某位同僚“觉醒1.0纯属发疯”的论调后当场哈哈大笑，仿佛全美民众都心照不宣地把那场闹剧当成了一场笑话；另一些人则忙着玩弄文字游戏重塑政治口号，言之凿凿地辩解称“停止资助警察”（defund the police）从来都不是真的要削减警务预算；还有一部分死硬派则负隅顽抗绝不退缩，至于他们究竟是历史的前锋还是撞上死胡同的末路狂徒，唯有时间能给出答案。但不争的事实是，在2025年1月特朗普开启第二个总统任期至翌年1月马姆达尼宣誓就职纽约市长的这一年间，美国左翼的价值轴心已然发生剧烈转向——从死磕种族与性别议题，全面回归到聚焦阶级矛盾本身。

回望那一时期互为因果的历史推力：无论是象牙塔学术圈的狭隘思潮向社交网络野蛮外溢，还是特朗普强势崛起带来的政治应激，抑或是“我也是”（#MeToo）反性骚扰运动的洗礼，以及2020年乔治·弗洛伊德惨死于白人警察膝下的悲剧——这些力量交织叠加，曾合力将美国左翼的视线死死焊死在反种族主义与反性别歧视的单一维度上。在那段狂热岁月里，身份认同被拔高为决定个人命运的终极标签：一个生活在贫困线以下的最底层白人，也被强行扣上享有哪怕最富有的非裔精英也未曾企及的“白人特权”。彼时的社会正义运动口头上高喊着重塑警民生态，实则彻底异化为美国多元精英阶层自娱自乐的精神内耗，诸如围绕常春藤盟校录取名额倾斜等脱离民生的议题争论不休。对于这群养尊处优的白领小资而言，四处挑刺种族特权沦为了炫耀自我道德优越感的通行证，鄙夷那些把选票投给特朗普的“底层可怜虫”同样成了时尚。任何一个受过基本政治经济学训练的正统马克思主义者，都能一眼看穿这是企图肢解分化无产阶级力量的经典拙劣把戏，然而民主党领导层当年却如获至宝般对这种撕裂社会的身份政治全盘照单全收。

失去的只有连任选票

随后，政治现实给出了严酷的当头棒喝。卡玛拉·哈里斯（Kamala Harris）在2024年大选中惨败收场，特朗普在非白人工人阶级选民中的得票率一路逆势飙升。紧接着，特朗普重返白宫后重拳整肃联邦多元、平等与包容（DEI）项目，最高法院亦挥舞法槌彻底废止了高校录取中的种族配额考量。与此同时，民主党基层活动人士的怒火瞬间转移到了加沙的苦难，原本紧绷的国内种族议题骤然失焦。更为决定性的一幕在于，马姆达尼正是凭借一套聚焦平抑生活成本、充满包容性的务实民生主张一路过关斩将爆冷胜选——弗吉尼亚州和新泽西州的温和派民主党州长候选人同样凭借这套民生王牌赢得了大选。

如今，尽管进步派与温和派依然龃龉不断，但双方在拯救挣扎求生的“不稳定无产者”（precariat）乃至传统产业工人阶级这一诉求上，达成了高度空前的默契。从佐治亚到阿拉斯加，今天的民主党人几乎全员摇身一变成为了经济民粹主义的拥趸，异口同声地呼吁提高最低时薪、压低恶性通胀并提供人人都负担得起的基本医疗保障。社会正义评判受害者与恶霸的标准，不再取决于当事人的肤色深浅，而完全取决于其银行账户余额的厚薄。亿万富豪被统一定性为吸血恶魔；而在民主党光谱极度宽泛的全新阶级界定中，凡是需要靠出卖劳动力打工糊口还贷的普通人，都被统战纳入了“工人阶级”的光荣范畴。这一战略转向最终究竟能否帮助民主党人走出低谷、乃至拯救在水深火热中挣扎的数千万美国劳苦大众，仍需拭目以待。但倘若这一天真的到来，这个百年老党至少应当由衷感谢那群曾饱受嘲弄的“年轻布尔什维克纨绔”——正是他们的折腾与反思，促使民主党猛然惊醒，重新找回了自己建党之初曾誓死捍卫的阶级立足之本。■

ch025.xhtml

美国宣称拥有委内瑞拉五分之一石油的开采权

该协议可能促使两国共同设法阻挠委内瑞拉的民主转型进程

图片来源：美联社

2026年8月29日 | 加拉加斯

这是一场充斥着浓郁特朗普个人色彩的重磅宣示。8月28日，美国总统高调宣布了一项被其冠以“世界历史上最大宗石油协议”的交易。他宣称，该协定将让美国获得对委内瑞拉逾650亿桶原油的“绝对控股控制权”，这一储量相当于该国已探明庞大石油储备的约五分之一。在讲话中，他极尽吹捧之能事，将民望极低的委内瑞拉实际掌权者德尔西·罗德里格斯（Delcy Rodríguez）尊称为“备受尊重的委内瑞拉临时总统”。罗德里格斯随即投桃报李，在肉麻的相互逢迎中，向唐纳德·特朗普致以“最由衷的谢意”，盛赞该协议将有力推动“本国经济全面腾飞”。

特朗普大肆夸口称，该协议使美国的战略石油储备“翻了一倍有余”。但这种说辞带有极大的误导性。拥有约460亿桶本土已探明石油储量的美国，绝非如特朗普所夸口的那样凭空被“赠予”了650亿桶委内瑞拉原油资产。这笔交易的实质是：美国政府借助五角大楼的渠道，直接注资收购了一家名为北美蓝色能源伙伴（NABEP）的私营跨国石油生产商母公司35%的控股股权。而委内瑞拉政权则投桃报李，将境内多达14处核心油田的开采特许权拱手授予了该公司，此外还附带了该公司此前已实际运营的另外3处油田。

据白宫方面披露的合作细则，美国政府将享有按开采边际成本优先平价承购NABEP未来总产量20%原油的核心特权，并对其余80%的产出拥有排他性第一优先采购权。不仅如此，华盛顿对NABEP董事会任何董事成员的任免均拥有一票否决权，且协议明文规定该董事会多数席位必须由美国公民出任。特朗普政府宣称该特许经营合同有效期长达整整100年，并有望在履约期内为委内瑞拉国库带来累计超2000亿美元的特许权使用费与税收红利；然而，委内瑞拉官方通报的协定合作期却仅为25年。

特朗普对这笔交易背后蕴含的浓厚帝国主义掠夺色彩毫不掩饰。“胜者理应通吃战利品，”他在9月1日接受采访时直言不讳。这番狂妄言论直指美军特种部队在今年1月3日发动的一场破门跨境奇袭——在那场闪电突袭行动中，美军在委内瑞拉首都加拉加斯强行生擒了强人独裁者尼古拉斯·马杜罗（Nicolás Maduro）及其夫人，并连夜将其押解包机引渡至美国本土羁押，二人目前正面临多项联邦跨国毒品走私罪名的刑事起诉。在特朗普描绘的战略蓝图中，一旦这场特种军事行动的“胜利果实”——委内瑞拉的大半石油血脉——被彻底纳入美国麾下，中东原油对美国的地缘战略牵制力将大大削弱，美国本土消费者的汽油零售价将被彻底打下来，而委内瑞拉自身也将“重获暴富”。

漫长的拉锯战

然而现实层面的阻碍要棘手得多。要想让委内瑞拉百孔千疮、设施残破的石油工业恢复元气并大幅提振产能，至少需要数年乃至更长时间的持续投入。在马杜罗及其前任乌戈·查韦斯（Hugo Chávez）两个无能政权长达二十余年管理混乱、贪腐横行的折腾下，这个曾经傲视全球的产油大国如今只剩下一堆锈迹斑斑的工业废墟。“在未来一年左右的时间里，委内瑞拉的实际增产能力在决定国际原油定价权的大棋局中根本无足轻重，”休斯敦莱斯大学贝克研究所委内瑞拉石油问题权威学者弗朗西斯科·莫纳尔迪（Francisco Monaldi）客观分析道。

更为致命的政治硬伤在于，特朗普此次并非在与一个代表合法民意的委内瑞拉国家签署契约，而是与该国未经民意洗礼的寡头集团做交易——以罗德里格斯为首的这批前朝遗老（其曾长期稳坐马杜罗政权的副总统要职），在权衡利弊后断然选择屈从于特朗普的威慑，以此换取自身免遭与其前主子一样身陷囹圄的悲惨下场。

图表来源：《经济学人》

在加拉加斯的大街小巷，平民百姓心头最深重的怨气并非控诉这项协议是美利坚的殖民主义巧取豪夺，而是痛恨它在客观上沦为了为非法政权续命输血的政治筹码——早在2024年的总统大选中，委内瑞拉民众就曾以压倒性多数选票将该统治集团逐出权力中心（然而那场大选在包括罗德里格斯在内的马杜罗核心内阁的暗箱操作下被悍然窃取）。“这一切全部都是背着全国老百姓在密室里私相授受，签字画押的全是一群通过舞弊窃据大位的非法掌权者，”正在加油站前排着望不到头长队等待加油的市民休玛拉（Xiomara）义愤填膺地控诉道。

特朗普挑中的商业代理人，更坐实了各方关于他正与被人民唾弃的残余寡头沆瀣一气的指责。NABEP背后的实际操盘控制人是亚历杭德罗·贝当古（Alejandro Betancourt），此人正是当年查韦斯执政时期借由腐败狂欢凭借内幕合同一夜暴富的委内瑞拉红顶商人之一。在委内瑞拉民间，他的名字因本世纪初查韦斯政府与其名下一家企业签署的一份天价电网改造维保大单而彻底声名狼藉。委内瑞拉反对派多年来严厉抨击该项目存在严重虚报工程款和交付缩水，指责正是当年的利益输送直接埋下了如今全境大停电频发的祸根。

捏着鼻子的交易

“贝当古恐怕是你能挑中的争议最大、名声最臭的商业合伙人了，”一名与委内瑞拉石油业界关系密切的知情人士坦言，“在公众认知里，他与这个国家烂到令人发指的电力系统彻底捆绑在一起。”尽管贝当古此前曾多次遭到欧洲司法机关涉嫌国际洗钱罪名的调查，但他始终全盘否认一切违法指控，且迄今未曾被任何一家法院定罪。一名美国高级官员为这项选择辩护称，贝当古是一个“深谙这门生意运营门道”的实干家，并着重强调其名下企业此前在已实际控制的3处油田上展现出了实打实的增产战绩。

美国官方言之凿凿地坚称，与NABEP达成的协议绝不会动用美国纳税人一分钱的财政公共资金。因此，筹措将原油日产量翻五倍拉升至100万桶所需的1000亿美元天量基建改造资金的重担，完全落在了这家企业自身的募资能力上。华盛顿打的如意算盘是：美国政府亲自入股背书的政治光环，足以给心存顾虑的全球潜在国际资本注入定心丸。

与此同时，特朗普政府正开足马力在各个层面大肆兜售“委内瑞拉正迎来绝佳投资窗口”的乐观论调。9月2日，特朗普任命的能源部长克里斯·赖特（Chris Wright）亲赴加拉加斯站台督师，见证了雪佛龙公司（Chevron）——在过去四分之一世纪的动荡岁月中唯一坚守该国未曾撤离的西方石油巨头——宣布拟在未来五年内追加投资70亿美元的重磅声明（该项目独立于前述NABEP协议之外）。然而，除雪佛龙之外，全球尚无任何一家跨国石油巨头跟进做出类似承诺。许多国际石油巨头依然对委内瑞拉心有余悸——2007年查韦斯政权曾强行推翻重写外资合资开采协议，直接导致多家国际油企损失数百亿美元、铩羽而归。

掣肘跨国资本脚步的另一个深层顾虑，在于该法律合同未来能否经得起政权更迭的合法性检验。倘若未来某天委内瑞拉通过真正透明公正的大选诞生了一个合法的民主政府，新政府是否会承认这份在霸凌高压下由各方公认的非法伪政权签署的不平等条约？“对于所有卷入其中的当事方而言，这注定将演变成一场巨大的国际闹剧与灾难，”委内瑞拉前计划部长里卡多·豪斯曼（Ricardo Hausmann）一针见血地警告道。

这随即引出了一种更为令人不寒而栗的严峻前景：鉴于未来的民主化选举将不可避免地对双方利益勾兑的温床构成致命颠覆，特朗普政府与当下的罗德里格斯伪政权之间，如今或许已滋生出了步调一致的共同利益诉求——那就是极力拖延委内瑞拉的民主转型进程。“随着该交易的全面铺开落实，反对派在国家政治版图中的话语权与存在感可能会遭到全方位的边缘化甚至无情扼杀，”曾长期常驻加拉加斯的前土耳其外交官伊姆达特·奥纳尔（Imdat Oner）深度剖析道。事实上，特朗普团队早已三番五次强力施压阻挠民望极高的委内瑞拉反对派领袖玛丽亚·科里娜·马查多（María Corina Machado）回国（她去年曾获准离境飞赴挪威领取诺贝尔和平奖）。面对这笔出卖国家核心资产的巨额暗箱协议，马查多至今保持着异寻常的沉默。

美国官方则义正辞严地对此予以坚决否认，坚称特朗普政府此举是在以大魄力托底委内瑞拉濒临崩溃的国民经济，为未来不可避免的真正大选奠定坚实的物质前提。美方辩解称，唯有确保下一届民选合法政府从一接手就能继承一个基础牢靠而非四分五裂濒临破产的经济基本盘，该国才能真正走出过去四分之一个世纪的人间浩劫。“我们的终极目标是扶植一个能够常态化、机制化举行自由大选的稳固民主共和国，”一名涉事美方高官信誓旦旦地宣称。

或许真能如其所愿吧。早在1975年，委内瑞拉时任能源矿业部长就曾一针见血地将石油痛斥为招致国家诅咒的“魔鬼排泄物”。回溯百年历史风云，凡是沾染这黏稠黑金利益的交易——更不必说那些包裹在宏大国家建构伪装下的地缘野心——似乎从来就未曾干净过哪怕分毫。■

ch026.xhtml

巴西当局试图封杀的总统候选人

雷南·桑托斯口齿伶俐、信奉自由主义，还组建了一支摇滚乐队，他以阿根廷总统哈维尔·米莱为榜样

在自由世界中摇滚 © Leonardo Carraro

2026年8月31日 | 里约热内卢

编者注（8月31日）： 本文发表后不久，兼任巴西高等选举法院法官的最高法院大法官若泽·安东尼奥·迪亚斯·托福利（José Antonio Dias Toffoli）暂停了雷南·桑托斯（Renan Santos）的候选人资格。托福利称，桑托斯的竞选团队未能正确备案其社交媒体账号。桑托斯的竞选团队表示将提出上诉

雷南·桑托斯自信爆棚。“我是同代人中世界上最出色的演讲家，”他在自己局促的办公室里这样宣称，语气半开玩笑半认真。他曾夸耀自己苦练瑞典语并达到流利水平，只是为了在一年一度的欧洲之旅中搭讪女孩，这趟旅行被他的朋友们戏称为“金发女郎之旅”。他贬斥对手“幼稚”、“可怜”且“平庸”。然而，尽管言辞浮夸，桑托斯却绝非骗子。这位42岁的候选人正凭借一份洋溢着政策极客（wonk）风格、描绘国家大胆愿景的纲领，角逐巴西下一任总统。

桑托斯的崛起与阿根廷言论激进的总统哈维尔·米莱（Javier Milei）颇有相似之处。同米莱一样，他也极度痴迷自由市场经济学。2014年，他创立了政治团体“自由巴西运动”（MBL为其葡语缩写），在一个长期推崇大政府传统的国家中倡导小政府理念。与米莱相似，桑托斯也将自己的教条与摇滚乐融合在一起，担任一支歌颂竞争优势和低税收的乐队主唱。（桑托斯自称音乐造诣更胜一筹；其助手则评价两位政客的乐队“烂得半斤八两”。）

此外，桑托斯同米莱一样是政坛局外人。他曾组织针对迪尔玛·罗塞夫（Dilma Rousseff）政府的大规模抗议活动，从而声名鹊起；这位左翼前总统于2016年被敌对的国会弹劾下台。直到2025年，MBL才以“使命”（Missão）为名注册为正式政党。该党目前在国会拥有一名议员吉姆·片桐（Kim Kataguiri）。除了桑托斯角逐总统宝座外，该党还将在10月的巴西大选中推举57名国会议员候选人。

步入“使命”党的办公室，会发现这是一个既专注于文化革命、又放眼于政治权力的工程。入口处散落着该运动的机关刊物《侍从杂志》（Revista Valete），上面曾刊登关于埃隆·马斯克（Elon Musk）的专题报道以及对酷儿女权主义哲学家朱迪斯·巴特勒（Judith Butler）的批判文章。这里的氛围更像初创科技公司，而非政党总部。程序员们在巨大的黑屏前飞快地敲击代码，走廊尽头则是一间用于拍摄节奏明快短视频的工作室，这些视频将被推送给该运动在网络上的400万追随者（以男性为主）。

桑托斯表示，聚焦媒体是对“觉醒主义”（wokeism）的回应。他解释称，取消文化的抬头“需要经典自由主义无法给出的回应”。他坦言，许多像他一样的千禧一代，曾经支持同性婚姻和毒品合法化，如今却“只因不愿使用某些代词……或被要求接受儿童可以改变性别，就被贴上法西斯的标签”。

为弱势群体呐喊 © Leonardo Carraro

但这嘲文化战“并非我们的核心身份”。桑托斯刻意将MBL与唐纳德·特朗普（Donald Trump）的“让美国再次伟大”（MAGA）运动拉开距离。经济议题依然是其最重要的主张。他表示，受美国自由意志主义者影响的“幼稚经济自由主义的失败”，促使该运动走向了新的思考。“认为只要国家退出所有活动并将一切私有化，问题就会迎刃而解，这是极其天真的想法。国家必须在许多领域发挥作用，尤其是在巴西这样一个贫富悬殊的国家。”该党目前认为，像巴西国家石油公司（Petrobras）这样的国有企业不应私有化，但应削弱政府对其的干预。负责该党经济方案的片桐将此称为“务实、非教条式的自由主义”。

桑托斯希望通过深层结构性改革，在30年内将巴西打造成世界强国。该党的“黄皮书”描绘了这一转型蓝图，这份长达350页的施政纲领吸取了世界各地成功的公共政策经验，力求应用于巴西。片桐曾专程前往丹麦研究其专利法，并赴日本考察地方政府评估体系。德国的双元制职业培训体系、中国的基础设施支出以及美国对人工智能监管的包容放任态度，都在其中有所体现。片桐表示：“我们绝不想要欧盟模式，他们沉迷于控制和监管，而不是发展自己的人工智能。”

其成果便是一套充斥着适用于巴西的新观念的政策方案。“使命”党提议将患者医疗记录数字化，并利用人工智能改进分诊流程，以缩短公立医院的排队时间。一项新法律将针对地方政客在教育、医疗和排污设施覆盖率方面设定绩效目标。达标官员将有资格从联邦预算中获得更多资金；而长期表现不佳者将面临暂时被剥夺任职资格的惩戒。人工智能还将用于提出精简公共部门岗位的建议。片桐滔滔不绝地赞叹道：“人工智能在提升行政效率方面的应用空间是无限的。”

直面核心重负

在削减开支的必要性上，桑托斯毫不避讳。巴西联邦预算中超过90%属于刚性支出，尤其是养老金。由于许多养老金和若干福利待遇与最低工资挂钩，而最低工资又随通胀和GDP增长自动上调，这部分支出的比重正持续无情扩大。到2029年，自由裁量支出预计仅占预算的0.1%，这几乎无法维持政府运转。“使命”党提议取消养老金与福利的挂钩机制，将教育和医疗支出与通货膨胀挂钩，而非与快速增长的财政收入挂钩。紧随其后的将是削减公务员的高额薪水以及推行养老金改革。片桐希望，所有这些举措能为拉动经济增长的基础设施投资腾出空间，他计划将基建支出占GDP的比重翻倍至4%。

谈及治安问题，民粹主义便压倒了自由主义。桑托斯的口号之一是“抓到就杀”（“Prendeu, Matou”），意指他认为警察应当如何对待犯罪分子。他屡屡称赞萨尔瓦多的政治强人纳伊布·布克尔（Nayib Bukele），后者通过将该国2%的成年人口投入监狱并践踏民主程序，大幅降低了犯罪率。（桑托斯自诩为“行事风格像布克尔，实质内核像米莱”。）他承诺将宣布紧急状态，并允许对黑帮展开一场“血洗”。“黄皮书”推崇“敌对刑法理论”（doctrine of the enemy），这一备受争议的法律理论认为某些罪犯不配享有公民权利。在实际操作中，执意打压政敌的行政部门极易滥用这一理论。

桑托斯在民调中的支持率一直徘徊在个位数，但他提出大胆构想的魄力正为他赢来巴西右翼重要人物的支持。许多人对弗拉维奥·博索纳罗（Flávio Bolsonaro）的参选感到灰心，后者是民粹主义前总统雅伊尔·博索纳罗（Jair Bolsonaro）的儿子，这位前总统因企图发动政变正在服刑27年。“弗拉维奥太软弱了，”一位桑托斯的支持者表示，“雷南博览群书、见多识广，能对众多不同议题侃侃而谈。弗拉维奥却一无所知。”桑托斯抨击博索纳罗家族在思想上平庸乏味且腐败堕落。

桑托斯将很难赢得女性选民的心。他曾公开称赞男性网络社区（manosphere）的博主。他最亲密的盟友之一阿图尔·杜瓦尔（Arthur do Val，曾于2022年与他一同前往乌克兰）曾扬言，乌克兰女性难民“很随便，因为她们穷”。在那次行程中，杜瓦尔吹嘘桑托斯曾向他传授如何搭讪女性的诀窍：“你必须去普通城镇……这些比较穷的小城镇最棒了。”

8月31日，兼任巴西高等选举法院法官的最高法院大法官若泽·安东尼奥·迪亚斯·托福利以桑托斯未能合规注册该党社交媒体账号为由，暂停了他的候选人资格。一天后，在舆论一片哗然之中，托福利撤回了决定。然而，这一被诸多专家视作极端的举措，反而强化了桑托斯的反建制信誉。即便他此次未能胜选，他也已经为那些心怀不满的右翼选民提供了一个替代博索纳罗主义（bolsonarismo）的新选择。■ # 对虾超级大国崛起

厄瓜多尔养殖户拥抱科技成效卓著

灰色黄金 © Reuters

2026年9月2日 | 瓜亚基尔

厄瓜多尔的对虾生产堪称精益制造的化身。水产养殖带的核心——瓜亚斯河三角洲上，大片长方形虾塘纵横交错。半咸水中隐匿着数以十亿计的半透明虾苗。幼体在温室般的孵化场中培育，在露天虾塘中育肥，被真空吸泵吸入卡车，随后运往包装厂进行冷冻处理。低效之处无时无刻不在被剔除。装满俗称“带头带壳”（HOSO）冻虾块的集装箱横渡大洋，主要自东向西穿过太平洋。“这就像一座工厂，”该国第二大生产商阿尔马尔集团（Grupo Almar）的沃尔夫冈·哈滕（Wolfgang Harten）说道。

图表：经济学人

这座工厂正在全负荷运转。出货量在过去十年中增长了三倍。2020年，厄瓜多尔超越印度成为全球最大的对虾出口国（见图表）。去年其对虾销售额突破80亿美元，超过了印度、越南和印度尼西亚的总和。曾经的石油之国，如今已成为“渔业国家”。对虾生产为该国约1.6%的人口提供了就业岗位，贡献了6%的GDP，这一比重甚至超过了汽车制造业对德国经济的贡献。该产业催生了该国最前沿的技术。运送对虾的货车均有武装护卫押送。对于沿海社区而言，在对虾加工厂谋得一份差事便是迈入中产阶级的门票。

生意并非一直如此顺遂。当厄瓜多尔在20世纪60年代开始养殖对虾时，它只是一项作坊式产业。凡纳滨对虾（Penaeus (Litopenaeus) vannamei），即人们熟知的白对虾或南美白对虾，原产于拉丁美洲。虾农从附近的红树林采集野生虾苗。鲜有人费心进行选择性繁育。丰厚的利润推动了行业扩张，却也带来了沉重代价。到2000年，乱砍滥伐导致该国红树林面积减半。白斑综合征在全球的爆发使许多虾农破产，染病虾塘中的对虾死亡率高达95%，该国产量在两年内锐减70%。

面对疫情，厄瓜多尔的亚洲竞争对手选择走集约化生产路线。池塘被塞得满满当当，密度高达每平方米300只虾。虾苗在实验室中进行隔离，以确保从未感染过白斑病毒。滥用抗生素成了家常便饭。厄瓜多尔却反其道而行之。生产商没有试图根除疾病，而是主动适应它。虾塘放养密度极低，仅为亚洲的十分之一。虾苗由白斑病幸存者繁育而来，这些幸存亲虾已具备抗病性。抗生素被逐步淘汰。

该国不断改良自身技术。自2010年代以来，一波又一波的创新浪潮让其养殖场跻身海洋科技的最前沿。亚洲的许多虾农仍每天两次划着独木舟抛洒一袋袋饲料。而在厄瓜多尔，养殖场配备了每三分钟投料一次的自动投饵机。水下麦克风通过录制甲壳动物咀嚼时发出的咔哒声，以判断它们是否吃饱。饲料制造商百佳（BioMar）的安德烈斯·里瓦杜拉（Andrés Rivadulla）表示，根据摄食速度，可以为下一个养殖周期量身定制配料。增氧机源源不断地向每个池塘输送氧气。机械过滤装置减少了水质污染。

其成效极为惊人。更健康的虾与更先进的技术相结合，意味着尽管单位面积产量低于泰国和印尼等地，但成活率却高得多。荷兰合作银行（Rabobank）的戈尔扬·尼科利克（Gorjan Nikolik）指出，厄瓜多尔约95%的对虾在捕捞时为活虾，而亚洲这一比率勉强达到80%。最大限度减少疫情爆发降低了成本，稳定了产量。培育本土物种也带来了优势。阿尔马尔集团每年能完成六个周期的50天养殖，周转速度几乎是印度和越南竞争对手的两倍。这促进了对虾企业的并购整合，并使家庭虾农融入资本雄厚的大型企业之中。

地缘政治亦助力了这片南美土地。厄瓜多尔已然是欧盟和中国最大的对虾供应国。然而，该国在北美市场长期难有作为，那里曾被印度的产能所主导。随后唐纳德·特朗普登场了。表面上因不满印度采购俄罗斯石油，他去年对印度商品征收了50%的关税。面对印度对虾近60%的综合关税（相比之下厄瓜多尔仅为10%），美国进口商纷纷倒戈。如今美国人消费的厄瓜多尔对虾超过了其他任何地区；即便今年早些时候印度的关税有所放宽，厄瓜多尔依然保持着领先地位。

切忌过犹不及

事实上，当前的主要问题在于产能过剩。自摆脱白斑病疫情恢复以来，养殖户的年产量以15%的速度递增。这使得市场供大于求，拖累价格跌至历史冰点。今年5月，厄瓜多尔的塘口收购价为每公斤2.40美元，相比之下，印度为2.99美元，中国为5.32美元。为了保护本国生产者，中国偶尔会以动植物检疫问题为由限制进口。巴西的进口关税高得令人望而却步。墨西哥则对走私怨声载道。此外还有气候变化，它实际上加剧了供求过剩。厄尔尼诺气候现象导致太平洋沿岸气温上升，虽不利于农作物，却极大地刺激了喜温对虾的生长。

因此，当务之急在于提振需求。巴尔干半岛和波斯湾被视作带头带壳冻虾的潜力市场。各大企业开始为美国消费者打造更高附加值的产品，例如去壳和裹粉对虾。快餐店正大力推广虾肉汉堡，营销人员则极力宣扬以虾肉为基础的酸橘汁腌鱼（ceviche）、天妇罗和意大利烩饭的健康益处。“开启崭新一天的最佳方式是吃一份对虾煎蛋卷，”厄瓜多尔国家水产养殖商会（National Chamber of Aquaculture）的何塞·安东尼奥·坎普萨诺（José Antonio Camposano）谈道。这套策略或许正在奏效。需求正在抬头。但厄瓜多尔的对虾巨头们仍渴望斩获更多。■ # 约韦里·穆塞韦尼离任后，乌干达将走向何方？

从近几个月的情况来看，前景绝难乐观

插画：Tarini Sharma

2026年9月3日 | 坎帕拉

在乌干达首都坎帕拉的大街小巷，两张张贴四处的巨幅海报格外醒目。一张海报上是一位身材发福的老人，戴着宽檐帽，目光冷淡严苛。另一张则是一位身着迷彩服、面容更加年轻而狂妄的男子。前者是81岁的总统约韦里·穆塞韦尼（Yoweri Museveni），他在40年前的一场内战后夺取政权，如今已垂垂老矣。后者则是穆塞韦尼喜怒无常的儿子穆胡齐·凯内鲁加巴（Muhoozi Kainerugaba）。

凯内鲁加巴将军现任乌干达国防军司令，但他对父亲的权位虎视眈眈，言行举止日益显露出仿佛自己已然掌权的姿态。年长的乌干达人表示，他的好斗与残暴让人联想起20世纪70年代声名狼藉的军事独裁者伊迪·阿明（Idi Amin）。他的野心正在军队、执政党乃至其家族内部掀起轩然大波，引发了人们对穆塞韦尼最终离任后政权内部爆发权力斗争的担忧。“我们的历史总在走极端：要么是无政府状态，要么是彻头彻尾的极权主义，”乌干达司法部长诺伯特·毛（Norbert Mao）指出。担心这一至暗时代卷土重来的人，绝非仅他一个。

穆塞韦尼于2024年任命凯内鲁加巴将军统领武装部队。然而，即便总统确实一直在将儿子培养为接班人，两人的性格却迥然相异。私下里，穆塞韦尼是一位富有魅力的演说家；而凯内鲁加巴将军往往寡言少语。父亲热衷于发表长篇大论；儿子却以在X平台上不得体的发泄而臭名昭著。“这几天我像猫一样清醒，”他在9月1日凌晨发帖写道。“为了安全起见，女性必须离我至少十米远……”

现年52岁的凯内鲁加巴将军在外界看来有时像个滑稽自负的“靠爹二代”（nepo-baby）。然而，乌干达人却鲜少笑得出来。在今年1月穆塞韦尼于一场充斥舞弊的选举中连任后的镇压行动中，这位将军曾扬言要杀掉并阉割主要反对派候选人博比·瓦恩（Bobi Wine）。瓦恩被迫潜伏，目前流亡海外。另一位著名反对派领袖基扎·贝西杰（Kizza Besigye）于2024年在肯尼亚首都内罗毕遭到绑架，此后一直被关押在狱中。他正以叛国罪受审；凯内鲁加巴将军发誓要将他处以绞刑。在7月的一次庭审中，贝西杰在发言中途倒地，随后被送医抢救。

近几个月来，这位将军的暴行变本加厉。今年7月，他的士兵绑架了坎帕拉前市长、贝西杰的辩护律师之一埃利亚斯·卢克瓦戈（Erias Lukwago）。这位陆军司令在X上发布了卢克瓦戈被蒙上双眼的照片，吹嘘他正身处自己的“地下室”中。卢克瓦戈随后被指控犯有叛国罪，目前仍身陷囹圄。据人权观察（Human Rights Watch）称，仅在6月中旬至7月中旬期间，安全部队就非法拘留了至少四名其他异见人士。据信其中数人遭受了酷刑。

在穆塞韦尼治下的乌干达，政治镇压并非新鲜事。但人们普遍感到，该政权正在蜕变为更加阴暗、更加专制的产物。国家机器曾试图用一层合宪性的薄纱来掩盖其暴行。“如今我们面临的是一个疯狂、毫无悔意、撕破脸皮的独裁政权，”去年被捕并遭受酷刑、如今同样流亡海外的律师埃隆·基扎（Eron Kiiza）这样说道。

持此类看法的不仅限于反对派。“我们在法律中所承诺的与实际发生的情况之间的鸿沟正在扩大，”司法部长毛坦承。他指出，军方剥夺了贝西杰在明显患病时获得医疗救治的权利，践踏了他的宪法权利。这位部长还收到了有关安全部队施加酷刑的可靠报告。然而，凯内鲁加巴将军“根本不在乎法律怎么写”，毛评论道。“我不知道我们是否已经倒退到了阿明时期——死尸漂浮在尼罗河上的程度，”另一位政府内部人士称，“我认为尚未至此。但这些都是令人极度担忧的严重问题。”

有什么力量能遏制这种滑坡吗？穆塞韦尼从未明确表示希望儿子继承大统。过去，他曾对其子部分极其出格的举动加以管束。但这位领袖近来日渐衰弱，对于约束儿子，他似乎日益感到力不从心或不愿插手。

这似乎助长了这位将军的嚣张气焰，促使他企图铲除眼中钉和竞争对手。今年5月，他将曾被视作动不得的实权人物、前议长安妮塔·阿蒙（Anita Among）软禁（目前已获释）。数名高级将领遭到清洗。8月25日，这位将军的爪牙绑架了奥德雷克·瓦布沃戈（Odrek Rwabwogo）的一位亲信；瓦布沃戈是穆塞韦尼的女婿，据信怀有问鼎总统宝座的野心。“这是我对奥德雷克那个蠢货的最后通牒，”将军在X上怒不可遏地痛斥，威胁要逮捕瓦布沃戈，并将该亲信活活打死或枪决。

唯一或许能让穆塞韦尼的朝堂恢复几分理智的人，是总统沉默寡言的弟弟萨利姆·萨利赫（Salim Saleh）。身为退役将军的萨利赫是一位幕后实权派。据了解，他对侄子的某些行径感到不安，被认为是除穆塞韦尼之外，凯内鲁加巴将军唯一不敢公开忤逆的人。

倘若家族世袭无法避免，乌干达的多数精英更倾向于由萨利赫出任总统。他被视作理性之人，在族群问题上态度相对温和。相比之下，凯内鲁加巴将军因狭隘的民族沙文主义而广受厌弃。他称自己的族群巴希马人（Bahima）为“特殊血统”，并用粗鄙的种族侮辱之词贬损其他族群。许多乌干达人对他与邻国卢旺达的密切关系也深表疑虑。这位将军常夸耀自己有两位叔叔：一位是萨利赫，另一位是卢旺达铁腕总统保罗·卡加梅（Paul Kagame）。“这两位叔叔，谁最终能对这位学生产生更大影响力？”前述政府内部人士反问道。

目前尚不清楚萨利赫的真实意图。有人认为他更愿意退居幕后充当“造王者”（kingmaker）。另一些人则认为，一旦政权内部爆发危机，他将出面干预，并在必要时亲自掌权。“‘劳尔·卡斯特罗式’的前景并非天方夜谭，”一位消息灵通的观察家表示。

凯内鲁加巴将军届时将作何反应，无人说得准。许多人推测，若局势到了不可收拾的地步，乌干达陆军的大多数力量将倒向萨利赫。然而，国境线对面的卡加梅叔叔，或许对局势该如何演变抱有自己的盘算。据说萨利赫与将军私交甚笃，令人难以想象双方会大动干戈。但在权力高度集中于少数人手中的国度，一场小小的家族内讧，就足以让整个国家分崩离析。■ # 埃塞俄比亚未公开宣称的无人机战争

对提格雷的一连串空袭将该国推向危险深渊

你们将听见打仗和打仗的风声 © Liu Fangqiang/Xinhua/Eyevine

2026年9月3日 | 内罗毕

在埃塞俄比亚最北部的提格雷省，8月标志着“阿申达节”（Ashenda）的序幕，这是一个东正教传统节日。这本该是家庭团聚、街头游行和音乐演出的欢聚时刻。然而今年，公共庆祝活动被悉数取消，私下的家庭聚会亦笼罩在压抑之中。

8月20日，埃塞俄比亚政府出动战斗机和无人机轰炸了提格雷首府默克莱（Mekelle）。据报道，空袭导致包括儿童在内的13人受伤。8月24日至8月27日期间，提格雷每天至少遭受一次轰炸。8月26日发生在南部的一次无人机袭击摧毁了一所学校。“他们企图恐吓并羞辱公众，”提格雷执政党“提格雷人民解放阵线”（TPLF，简称提人阵）的一名高级官员表示，“反抗是唯一的出路。”

地图：经济学人

局势升级绝非朝夕之变。2020年至2022年间，埃塞俄比亚政府与提人阵爆发了一场残酷内战，据信夺走了数十万人的生命。平息战火的和平协议并未消除埃塞俄比亚总理阿比·艾哈迈德（Abiy Ahmed）与提人阵领导层之间的核心矛盾；在阿比于2018年将他们排挤出核心层之前，提人阵曾主导埃塞俄比亚政局数十年之久。最近数月，埃塞俄比亚政府军与效忠提人阵的部队爆发了多次小规模交火。在许多观察家看来，重返内战早已近乎不可避免。

然而，近期的袭击也令人感到匪夷所思。与上一场战争不同，埃塞俄比亚政府并未切断提格雷的电信通讯。埃塞俄比亚首都亚的斯亚贝巴与默克莱之间的民航商业航班仍在执飞。2020年，阿比在国家电视台向提人阵宣战后才派出坦克开进该地区。而这一次，他的政府甚至从未对无人机袭击发表过公开评论。相反，数月来悬而未决的紧张对峙悄然演变成了一场实质上未正式宣战的闪电空袭。提格雷如今沦为一片灰色地带：不再处于和平之中，却也尚未全面开战。

陷入这种模糊状态的部分原因，可能是阿比这次行事更为审慎。他在2020年的好战言论以及埃塞俄比亚军队犯下的一系列暴行，玷污了他曾经的和平使者形象。若因挑起另一场战争而遭受指责，来自国际货币基金组织（IMF）和世界银行的关键经济援助可能化为泡影。所幸对阿比而言，提人阵竭力展现出侵略者的姿态。数月来，该组织一直强征提格雷青年入伍，并扬言要将阿比政权“埋葬”。

阿比不愿贸然出击的另一个原因在于，地区结盟关系发生了翻天覆地的变化。在前一次战争中，来自提格雷南部阿姆哈拉地区的民兵组织和其北部厄立特里亚的军队曾与阿比并肩对抗提人阵。如今，两者均与往日的死敌达成了不可思议的联盟。尤其是厄立特里亚，已成为阿比的核心死敌。尽管该国老牌独裁者伊萨亚斯·阿费沃尔基（Isaias Afwerki）对提人阵深恶痛绝，但他现在更忌惮这位埃塞俄比亚总理吞并该国部分海岸线的企图。“厄立特里亚人可不傻，”提人阵官员指出，“他们将亚的斯亚贝巴政权视为生死存亡的威胁。”

交战双方似乎都对和谈缺乏兴致。提人阵声称结束上一场战争的协议已“名存实亡”。埃塞俄比亚官方媒体则宣称提人阵已“不复存在”。8月24日，提人阵主席勉强会晤了尼日利亚前总统奥卢塞贡·奥巴桑乔（Olusegun Obasanjo）——后者是该垂死协议的调解人之一。提人阵表示，会谈“既未达成新方案，也未展现任何实质性进展”。

局势将走向何方？由于不愿在多条战线上同时与多方强敌展开地面决战，阿比很可能将重点放在争夺提格雷的制空权上。埃塞俄比亚军队对提人阵拥有近乎绝对的空中优势。在与阿姆哈拉叛乱分子的交战中，政府军严重依赖空中轰炸，导致大批平民丧生。去年，东戈贾姆发生的一起无人机空袭造成包括一名8岁男童在内的16人死亡。此类事件在阿姆哈拉已屡见不鲜，甚至鲜少引起媒体关注。

提人阵方面则发誓要夺回西提格雷——这片肥沃的土地在上一场战争中被与阿比结盟的民兵实施了种族清洗。然而，由于多年内讧元气大伤，外界普遍怀疑其是否尚有此等能耐。自2025年初以来，已有数名政治领袖和军事指挥官倒戈投向阿比。这无疑为埃塞俄比亚政府提供了大量关于提人阵军事基础设施和行动计划的情报。一些观察家认为，连绵不断的无人机袭击已使该组织陷入群龙无首的混乱。“他们太虚弱了，根本无力反击，”亚的斯亚贝巴的一名分析师评价道。

事态的发展很大程度上取决于提人阵能从厄立特里亚获得多少支援——鉴于该组织与伊萨亚斯宿怨深厚，这仍属未知数。它或许还在期盼苏丹武装部队（SAF，苏丹国家正规军）的跨境援助，苏丹军方对阿比在苏丹内战中支持其死敌“快速支援部队”怀恨在心。

8月1日，提人阵与埃塞俄比亚军队在谢雷里纳（Shererina）爆发冲突，该地是西提格雷靠近苏丹边境的一处狭长地带。一家与埃塞俄比亚政府关系密切的智库随后指责苏丹武装部队援助提人阵。上周，苏丹武装部队表示击落了从埃塞俄比亚境内发射的三架敌方无人机。提人阵领导层坚称，厄立特里亚和苏丹武装部队誓将阿比赶下台。提格雷这片灰色地带的危机，随时可能将整个地区染成一片血海。■ # 伊朗代理人走上各自道路

经过三年的战争，“抵抗之弧”不再浑然一体

摄影：Getty Images

2026年9月3日 | 迪拜

多年来，伊朗官员口中总挂着“抵抗之弧”（axis of resistance），指代从贝鲁特到萨那的广大盟友网络，他们本该同仇敌忾，抗击共同的敌人（见地图）。然而，当伊朗在过去15个月内两度遭到袭击时，各方对是否参战却作出了大相径庭的抉择。这个同盟或许并未消亡，但其品牌号召力已经超越了实质同盟的存续能力。

地图：经济学人

扶植阿拉伯民兵武装的战略可追溯至20世纪80年代，当时伊朗派遣伊斯兰革命卫队（IRGC）前往黎巴嫩训练真主党武装人员。在伊朗眼中，此类组织可作为其疲软常规武装力量的替代品：它们充当前沿防线，对外投射力量，同时将战火隔绝于伊朗国门之外。

这一防线显然失灵了。2023年10月7日由哈马斯主导的血腥屠杀引发了连环战争，最终导致伊朗本土遇袭。真主党和哈马斯遭受沉重打击。叙利亚的巴沙尔·阿萨德（Bashar al-Assad）政权已然垮台。伊拉克政府则勒令各民兵组织在10月前缴械。

该政权的一些敌对势力期盼伊朗被迫收缩支援（或许借此促成与美国的协议），进而触发整个同盟的解体。但这不太可能发生。民兵武装虽身负重伤却仍具生命力，且支持他们的战略逻辑比以往任何时候都更加充分。然而，该同盟若想重组为一个联合集团也将举步维艰：位于德黑兰的母公司已囊中羞涩，其名下的诸多子公司亦经营不善。

更有可能的结果是，其各个组成部分虽得以残存，却渐行渐远。一部分将沦为伊朗的附庸，虽被牢牢掌控却鲜少动用；少数将彻底脱离这一阵营；另一些则将沦为自谋出路的散户：这些代理人可能会另寻归宿。反直觉的是，在自身能力遭受最毁灭性打击的组织身上，伊朗反倒施加了最严密的控制。

以真主党为例，它曾是伊朗最为令人生畏的代理人。直至2023年，它仍坐拥超过10万枚对准以色列的火箭弹与导弹，以及一支训练有素、在保卫阿萨德政权的战火中磨砺出过硬本领的突击部队。其魅力非凡的领导人哈桑·纳斯鲁拉（Hassan Nasrallah，于2024年被以色列刺杀）视自己为平起平坐的伙伴而非受人摆布的棋子。他是深受伊朗革命意识形态浸润的虔诚信徒，却始终竭力维系一定的自主权。

今非昔比。纳斯鲁拉的继任者纳伊姆·卡西姆（Naim Qassem）不过是个不得人心的平庸主管。以色列的定点清除夺走了该组织大量经验丰富的高级指挥官。黎巴嫩政府坚决主张真主党必须交出其大幅缩水的军火库。由于指挥层死伤殆尽，伊朗革命卫队直接派遣己方人员进驻黎巴嫩实施掌控，这让真主党别无选择，只能在今年3月挑起与以色列的新一轮战火。与此同时，该组织在黎巴嫩国内的民望已然跌落谷底。伊朗之所以力保真主党，更多是不愿承受其覆灭所释放的政治信号，而非看重其实际作战能力。

与之形成鲜明对比的是胡塞武装——伊朗盟友中最具独立自主意识的一支。他们在20世纪90年代起家于也门北部的本土叛乱。数十年后，伊朗才向其提供大规模资金与军事援助。而援助一旦到位，便令其作战能力突飞猛进。曾被讥讽为赤脚枪手的胡塞武装分子，如今操纵反舰导弹在红海袭扰过往船只，运用防空系统击落美国的“死神”（Reaper）无人机。

摄影：Getty Images

然而，胡塞武装始终极力捍卫自身独立性。在10月7日之后的连环战争中，他们的介入时紧时松，其行动更多基于自身利益考量而非听命于德黑兰。他们自主生产武器的能力与日俱增，并正开辟自身的对外结盟，例如与控制索马里大片领土的基地组织分支“青年党”（al-Shabab）结盟。胡塞代表团曾访问俄罗斯，并与中国展开对话。伊朗与胡塞武装仍将保持共同利益——双方皆敌视美国、以色列和沙特阿拉伯。但这与唯命是从有着本质区别。

介于两者之间的则是那些离心离德的散兵游勇。伊拉克的许多什叶派民兵组织是2003年美国主导的伊拉克战争的副产物；战后，伊朗源源不断地输送武器与资金，以期借机消耗美国宿敌。二十年后，这些民兵的优先诉求已然改变。伊朗固然仍是盟友，但他们最大的金主已变为伊拉克政府，后者供养着超过20万名领取国家薪饷的民兵武装人员。

部分民兵今年早些时候确曾对美军基地发动过袭击，但他们担心与伊拉克政府或其盟友爆发公开冲突会危及自身的既得利益。据称伊朗革命卫队已派遣顾问赴伊拉克建立小型战斗小组，专门负责袭击海湾国家，这一迹象表明各大主要民兵组织并不情愿卷入其中。

哈马斯在这一同盟中历来显得格格不入——它是一个身处什叶派为主集团内的逊尼派组织。它曾因伊朗支持叙利亚阿萨德政权而与德黑兰反目。其新任命的领导人哈利勒·哈亚（Khalil al-Hayya）虽被视作伊朗的忠诚盟友，但其他高层更倾向于向土耳其及海湾等逊尼派大国靠拢。

今年3月，哈马斯敦促伊朗停止对海湾国家的袭击。伊朗不会彻底抛弃哈马斯——支持巴勒斯坦人是德黑兰政权合法性的核心所在——但其支持可能更多停留在口头声援而非物质输送。

即便伊朗决意重整麾下所有盟友，其资源也已捉襟见肘。战争造成了数千亿美元的惨重损失。美国正封锁其港口并加码经济制裁。阿萨德政权的垮台亦切断了通往黎巴嫩的陆路走廊。伊朗必须分清轻重缓急。

从某些维度审视，该同盟的宿命或将折射出20世纪80年代苏联卫星国的轨迹——彼时的莫斯科已无力再提供慷慨援助。诚然，这一类比并不完全贴切。莫斯科的盟友当时尚有另一敌对超级大国可以依附，而德黑兰的走卒们面对的则是一盘更为凌乱的棋局：部分组织拥有诸多潜在伙伴，另一些则彻底孤立无援。然而不可否认的是，中东这间野心勃勃的“托拉斯”，或许正不可逆转地面临一场资产重组。■ # 议会选举前夕，本雅明·内塔尼亚胡挑起内耗

竞选活动步入崭新阶段

摄影：Reuters

2026年8月31日 | 耶路撒冷

今年夏天，“特洛伊木马”的身影似乎无处不在。在针对10月27日以色列议会选举的竞选宣传片中，以色列总理本雅明·内塔尼亚胡（Binyamin Netanyahu）便将其作为核心道具。在广告中，目前在民调中遥遥领先的前军方将领加迪·埃森科特（Gadi Eisenkot）以AI生成的形象骑在木马上。而潜伏在木马腹中的，则是另外两位反对派领袖：曼苏尔·阿巴斯（Mansour Abbas）与亚伊尔·戈兰（Yair Golan）。

内塔尼亚胡传递的信息粗暴而明确：埃森科特唯有联合阿巴斯领导的以色列阿拉伯联合名单党（Ra’am）以及戈兰领导的犹太复国主义左翼民主党，才可能胜选。在《经济学人》的民调追踪中，埃森科特领导的中右翼政党“正直党”（Yashar）支持率已反超内塔尼亚胡的利库德集团（Likud）。但与内塔尼亚胡一样，他必须拉拢形形色色的政治伙伴方能组建政府。

埃森科特在利库德基本盘——中右翼选民中与日俱增的声望，令内塔尼亚胡阵营坐立不安。因此，利库德集团试图通过关联抹黑来毁其声誉。污名化代表以色列20%以上人口的阿拉伯政党是一贯套路。内塔尼亚胡曾反复给他们扣上“恐怖主义支持者”的帽子，指责联合名单党是穆斯林兄弟会的附庸（尽管在走投无路时，他也并不排斥寻求阿巴斯的支持）。

然而，将同一招数施加于犹太复国主义政党身上尚属首次。十年前，时任以色列国防军副总参谋长的戈兰曾将以色列当时的社会风气比作大屠杀前“欧洲发生的令人毛骨悚然的进程”。在最近的加沙战争期间，他又直言：“一个健全的国家不会针对平民发动战争，不会将杀戮婴儿当作消遣嗜好，更不会将驱逐平民确立为既定目标。”

戈兰的政敌如今正利用这些言论，将他的政党描绘为激进分子、非犹太复国主义者，因而不具备执政资格。但迄今为止，民调走势并未倒向内塔尼亚胡，且各大主要反对党无一明确排除与民主党联合执政的可能性。

然而，在下届政府是否吸纳阿拉伯政党的问题上，利库德集团成功分裂了反对派阵营。戈兰是唯一公开表示支持的人。他所有潜在的合作伙伴要么坚决反对与阿拉伯政党共事，要么言辞闪烁。在公开场合，埃森科特表示愿考虑与任何接受一套笼统原则的政党联合执政。但相比吸纳以色列的阿拉伯代表，他宁可领导一个弱不禁风的少数派政府。

尽管内塔尼亚胡的极右翼与宗教政党联盟或许无力赢得议会多数，但他已然开始主导竞争对手组阁谈判的基调与规则。他的选战攻势正将埃森科特推向更加强硬的民族主义立场；后者近期已排除了在执政期间承认巴勒斯坦国地位的可能性。即便埃森科特成功组阁，恐怕也只能勉力维系一个寿命短暂的内阁，其推行的政策与即将卸任的本届政府并无二致。

身为以色列政坛最精明的政客之一，阿巴斯是迄今为止唯一加入过执政联盟的阿拉伯裔以色列领袖。2021年，他曾力挺由右翼政客纳夫塔利·贝内特（Naftali Bennett，今年再度角逐）领衔的政府。阿巴斯声称，那些以联合名单共同参选的其他阿拉伯政党拒绝作出入阁推行变革所需的妥协。

为了提升再次入阁的胜算，联合名单党今年将首次推举犹太候选人。对一个由保守派伊斯兰主义者主导的政党而言，约阿夫·塞加洛维奇（Yoav Segalovitz）的入选颇为出人意料。作为一名世俗派中间人士，塞加洛维奇最大的资本是他昔日的警务调查官履历。阿巴斯的首要任务之一，便是促使目前由极右翼犹太力量党（Jewish Power）党魁伊塔马尔·本-格维尔（Itamar Ben-Gvir）统领的警方，严肃对待荼毒以色列阿拉伯人社区的血腥黑帮暴力潮。

与此同时，阿巴斯也希望借此降低犹太政党对联合名单党的排斥感。塞加洛维奇或许能助一臂之力。与内塔尼亚胡如出一辙，阿巴斯不仅谋求选票，更意在重塑执政联盟的构建规则。阿巴斯不愿沦为蜷缩在特洛伊木马腹中的隐匿看客，他渴望成为驾驭木马的执缰之人。■ # 德国选择党将如何施政

若夺下首个联邦州，整肃校园与公共传媒的意识形态仅是冰山一角

摄影：HANNIBAL HANSCHKE/EPA/Shutterstock

2026年8月31日 | 马格德堡与策尔布斯特

在德国东部萨克森-安哈尔特州小镇策尔布斯特的集市广场上，午后酷热难耐，乌尔里希·齐格蒙德（Ulrich Siegmund）脸上的笑容却未曾褪去分毫。整整三个小时，作为民粹右翼德国选择党（AfD）在9月6日萨安州选举中的领衔候选人，齐格蒙德（见图）顶着烈日为苦等良久的狂热支持者们在T恤上签名并合影。以往选择党集会上常见的戾气在此荡然无存。“这不仅是一个政党，这是一场运动！”当地选择党新秀菲利普-安德斯·劳（Phillipp-Anders Rau）喜笑颜开。

图表：经济学人

打出“一切皆有可能！”这一欢快口号的齐格蒙德，期盼成为二战后德国首位极右民族主义联邦州长。在周日的选举中，选择党势将击败执政的基民盟（CDU）并斩获大胜（见图表）。取决于跨过5%得票率门槛的政党数量，该党甚至可能独立斩获议会过半席位。即便票数略有不足，齐格蒙德亦有其他通往权力宝座的途径，例如策动基民盟议员倒戈——尽管他向本刊特派记者坚称自己对此毫无兴致。

人群中的支持者们对该党包藏祸心的指控不屑一顾。“选择党并不激进，它只是代表言论自由，”当地居民海克·施密特（Heike Schmidt）表示。然而，在其长达258页的执政纲领中为萨安州描绘的蓝图，是有望执政的政党所炮制出的最激进构想之一。这无异于勾勒出一种新型国家的框架，党内高层视其为全德其他各州的施政样板。但这些政策既缺乏财政预算支撑，在某些领域亦显得荒诞不经。

部分主张（诸如废除庇护权或恢复与俄罗斯的关系）本就不属于州政府的职权范畴。但属于其管辖权力的政策同样繁多。首先是教育。选择党主张将残障儿童和难民子女剥离出普通课堂，专家指出这公然践踏了欧洲法和国际法。该党谋求增设俄语教学，恢复与俄罗斯学校的交流机制。其目的还在于废除强制义务教育法（Schulpflicht），以便心存疑虑的家长能够在家对子女进行家庭教育。选择党计划重写教学大纲，剔除“觉醒主义”内容，颂扬德国历史成就。该州党部首席理论家、纲领的主要起草人汉斯-托马斯·蒂尔施奈德（Hans-Thomas Tillschneider）辩称，德国对纳粹历史的自我反思文化导致了全民性的“认同障碍”。“多一点俾斯麦，少一点希特勒”便是其开出的药方。

选择党对德国现存体制下的大学深恶痛绝。蒂尔施奈德曾将人文社科学术界辱骂为社会变革的“合法化娼妓”。该党谋求重塑大学治理架构，退出现行的欧洲学位授予标准体系，并设立人口学教席以及专门的“批判性伊斯兰研究”所。由于学术自由受到宪法明文保护，该党将很难推行上述绝大部分构想。但鉴于大学资金主要来自各州财政，一个态度强硬的政府完全可以通过收紧财务枷锁来施加压力。

齐格蒙德已承诺取缔长期被德国民粹右翼视作眼中钉的公共广播机构，代之以“中立”的替代媒体。选择党的施政纲领将矛头直指萨安州“拥抱现代”（#moderndenken）的宣传口号，以及该州对发源于德绍的20世纪包豪斯建筑运动的推崇。在该党眼中，包豪斯恰恰代表了其极力斩草除根的无根文化。该党给出的替代方案是“思我德意志”（#deutschdenken），着力强调马丁·路德、弗里德里希·尼采和神圣罗马帝国的历史功勋，并提出推行“爱国主义旅游概念”。

然而，落实所有这些主张所需展现的治理能力，即便在选择党内部也饱受质疑。治理德国联邦州需要成百上千名官员与行政人员，而除了执掌过区区几个城镇与县级行政区外，选择党毫无实际施政经验。齐格蒙德最亲密的副手中有两人曾是新纳粹青年组织的成员；这绝非治理一个邦州的理想人选。去年成立的“黑红金学院”正着手在行政管理方面培训党内骨干，选择党亦向曾参与奥地利州级和国家级联合政府的民粹右翼自由党取经。但就应对眼前艰巨的任务而言，这些准备显然杯水车薪。

哈勒大学的佩特拉·多布纳（Petra Dobner）指出，财政瓶颈同样将束缚选择党的手脚。萨安州是一个拥有210万人口的穷困州，背负着约250亿欧元（约合290亿美元）的巨额债务，高度依赖德国其他联邦州的财政转移支付。根据哈勒经济研究所的测算，选择党的执政纲领包含对托幼服务和免费校园餐食的慷慨许诺，每年将产生高达22亿欧元的财政缺口。或许是意识到了这一点，选择党近期出炉的“百日施政纲领”主要聚焦于悬挂德国国旗、废除性别中立语言等象征性议题。

然而，处于该党攻击火力网内的各类机构并未掉以轻心。数月以来，各大学、法官协会及非政府组织一直在制定应对选择党执政的防范战略。今年早些时候，萨安州议员修改了推举法官及议长的程序规则。人们普遍担心的是，选择党可能会利用法院需要数年时间方能推翻违宪举措的司法空窗期，强行推行无数违宪法案，以此瓦解政敌的抵抗意志。

选择党执政的前景正促使德国其他地区高度警觉。鉴于该党与俄罗斯暗通款曲，其余15个州的内政部长一直在闭门商讨如何将其排除在跨部门情报协调机制之外。图林根州内政部长格奥尔格·迈尔（Georg Maier）赞成构筑“防火墙”（Chinese walls），确保重大国家安全机密不被该党提名的部长所接触。

选择党同样忌惮萨安州内部“深层政府”的掣肘。齐格蒙德放言将寻求撤换多达200名州级公职人员，并威胁对阻碍其政策落实的人员采取“行政惩戒措施”。该党可以推行极少数政治任命，并有权解雇若干高级官员。他们势必会罢免该州情报机构负责人，后者曾于2023年正式将萨安州选择党定性为“极右翼极端主义组织”。由于大多数普通公务员依法不可被随意解雇，关注该州司法韧性的法学学者弗里德里希·齐勒森（Friedrich Zillessen）指出，选择党政府可能另辟蹊径：借助公务员退休进行人员自然减员、提拔亲近人士，或刻意制造恐吓氛围。

萨安州的社会现状难言健康。它是德国各州中老龄化最严重、贫困率最高的地带之一。选择党誓言通过感召离乡的德国本土人才回流，以此替换掉“文化异己”的专业人才。但其政策导向极有可能适得其反。坊间充斥着若选择党上台大批劳工便将逃离的传闻。原本可能考虑迁居此地的外来人口，尤其是外国技术人才，势必会望而却步。该党甚至扬言拒绝接收总理弗里德里希·默茨（Friedrich Merz）去年设立的国家基建基金中归属该州的份额。“如果我们都因为忍无可忍而离去，这里最终会变成什么样？”当地一家企业的管理人员索菲·容尼克尔（Sophie Jungnickel）发问道。

不少选择党的政敌似乎倾向于任由齐格蒙德自行把事情搞砸，寄望于一个穷乡小州的执政挫败能够打破该党看似不可战胜的光环。萨安州当地居民对此却另有感触。“坐在柏林或德国南部，说这些风凉话自然轻巧，”近期刚搬离该州的政治活动家莱纳特·伯特（Lennart Birth）直言，“但对于每天必须承受这一切的人来说，绝非易事。”■ # 困兽犹斗：弗拉基米尔·普京意图升级战争

他拒绝退让的执念根植于俄罗斯的黑帮政治逻辑

摄影：IMAGO

2026年8月30日

8月29日，弗拉基米尔·普京在莫斯科的一所大学指出，虎鲸以北极熊为食。“这就是食物链，”这位俄罗斯总统宣称，“应当把这些教给孩子们，让他们明白我们生活在一个怎样的世界，以及我们为何要开展特别军事行动”——即俄罗斯对乌克兰的战争。这种说法完全是无稽之谈：自然界中至今没有虎鲸捕食北极熊的确凿记录。但它却生动展现了这位俄罗斯总统弱肉强食的丛林世界观。

近来，北约盟国对普京的危险思潮日益感到忧心忡忡。多年来，俄军在战场上整体陷于僵局。这场战争已演变为一场无人机与导弹对攻的拉锯战，虽然重创了乌克兰，却也让俄罗斯自身千疮百孔。但普京绝无退缩之意。令人担忧的是，他可能会像过去那样，通过寻找升级冲突的新手段来打破死局（tupik）。

8月25日，美国中央情报局（CIA）局长约翰·拉特克利夫（John Ratcliffe）突访莫斯科。关于此行目的众说纷纭，有人称其旨在警告普京切勿袭击北约国家，也有人认为他是去传达美方对俄前景的严峻评估以力劝其开启和谈。三天后，芬兰请求瑞典派遣战机与军舰协同监控芬俄边境的越境渗透。9月1日，德国公开指责俄罗斯对8月4日莱比锡机场遭遇的无人机袭击负有责任，并象征性地追加了微不足道的制裁。8月下旬，罗马尼亚击毁了一架瞄准其水域钻井平台的俄罗斯无人艇，斯洛伐克则挫败了一起针对无人机制造厂的纵火阴谋。

与此同时，普京打破了对乌克兰无人机袭击俄罗斯炼油厂和电商仓库一事的沉默。在8月22日的讲话中，他指责乌克兰打开了“潘多拉魔盒”，并放言要“对你们经济中最脆弱的命脉予以反噬打击”。当天早些时候，他粗暴驳斥了任何有关终结战争的构想：“特别军事行动仍在继续。”

克里姆林宫核心圈消息人士透露，在整个夏季经过数周盘算后，他最终决定升级战事。事实上，一位与普京交往甚密的人士透露，他极可能将拉特克利夫的到访视作美国软弱怯懦的信号。该消息人士称，普京当年便是以同样的逻辑去解读乔·拜登（Joe Biden）的CIA局长比尔·伯恩斯（Bill Burns）于2021年11月赴莫斯科警告其切勿入侵乌克兰的行程。

无人能确切窥探普京的内心活动。但他升级战事的动机同当初决定入侵如出一辙，均源于内部考量。他的核心诉求自始至终都是保全自身权力，并最终保全自身性命。

令克里姆林宫深感忧虑的，是公众对战争的默许态度正在分崩离析。乌克兰无人机的纵深打击，叠加网络的封锁管制，彻底击碎了正常生活的虚幻表象。独立民调机构列瓦达中心（Levada Centre）的数据显示，57%的俄罗斯受访者用“焦虑紧绷”来形容当下的局势；受访者将此归咎于战事蔓延、无人机突袭以及未来前景的迷茫。普京的民意支持率正在持续下滑。

精英阶层的态度更为低迷消极。一位内部人士直言，最大的痛点在于这场战争所耗费的时间成本：无论最终结局如何，打了近五年本身就是一种惨败。另一位消息人士透露，越来越多的人开始意识到“沙皇已然一丝不挂”。俄罗斯目前暴露出的问题虽不足以对普京构成致命威胁——他坐拥充足的经济资源和国家镇压机器，足以将战争维系下去。然而，外界的观感与认知至关重要。

普京或许善于用大国地缘政治的华丽辞藻来粉饰其战略意图，但他在本质上始终奉行着黑帮生存法则：绝不暴露软弱。俄罗斯的体制对暴力具备极高的容忍度，却对失败绝不姑息。他深知，哪怕无法达成全部拿下乌克兰东部顿巴斯地区的最低战役目标，都将令其自身陷入万劫不复的险境。（据一位知情人士透露，在战争期间的某一时刻，他曾对密友直言：“他们会把我吊死。”）单纯维持现状绝非破局之策。“被动维持局面的脚本有害无益，”一位俄罗斯精英高层指出：这只会不断透支经济资源并耗尽政治资本。

目前尚不清楚俄罗斯将通过何种具体方式升级冲突。它极有可能加剧针对乌克兰盟友的混合战争，掐断西方对乌克兰的后勤补给通道。西方情报机构担忧普京可能会将矛头对准欧洲本土的援乌军工厂。但他们研判，普京目前仍极力避免与北约爆发正面冲突。

就乌克兰战场内部而言，局势升级在很大程度上意味着对既有手段的大肆加码。8月30日，俄罗斯国防部宣称正酝酿对乌克兰能源基础设施发动超大规模打击。然而乌克兰方面对此类打击早有防备。至于地面战场，俄罗斯内部消息称普京执意再向“绞肉机”战场填入30万至40万兵力。新增兵源可能交由各地方政府自行征募，以期将政治不满情绪封锁在地方层面。当局已在莫斯科东南750公里的奔萨州先行试点推行全新的征兵体系。

普京同时在加码对所谓“内部敌人”的镇压清洗，这并非出于防范实质威胁，而是为了向外界展示强权肌肉。9月2日，警方逮捕了重症儿童临终关怀医院的创始人丽迪娅·莫尼亚瓦（Lidia Moniava）。她因2023年在社交媒体上发表的一条反战言论，面临最高10年的监禁刑期。而就在两天前，当局以身体抱病为由，撤销了对圣彼得堡富商伊利亚·特拉伯（Ilya Traber）的所有指控。这位与普京交情匪浅的商界寡头此前曾因涉嫌雇凶杀人而遭到指控。■ # 哈康八世登基：挪威新王任重道远

其先父深受保皇派与共和派的共同爱戴

摄影：Reuters

2026年9月3日

挪威向来不崇尚繁文缛节。但在哈康八世（Haakon VIII）于9月1日在议会庄严宣誓效忠之际，适当的隆重排场仍不可或缺。新国王乘坐着其父亲哈拉尔五世（Harald V）在1968年大婚时使用的林肯大陆（Lincoln Continental）轿车抵达现场。哈康身着军礼服，在王储英格丽德·亚历山德拉（Ingrid Alexandra）的陪同下，郑重誓言将继承父亲的遗志，维系王室与议会之间的紧密融洽关系。议员们伴随着《天佑吾王》的旋律，齐声高唱挪威王室颂歌《国王之歌》（Kongesangen）。

尽管挪威人崇尚人人平等的社会风尚，但公众对君主制普遍持拥护态度。不同寻常的是，这一政体曾获得全民公投的背书：1905年挪威与瑞典解除联邦条约后，79%的选民通过投票选择了拥立国王而非建立共和国。哈拉尔于8月28日以89岁高龄驾崩后，诺斯塔特（Norstat）民调显示，仍有72%的挪威人支持保留君主制。

这在很大程度上要归功于哈拉尔的个人操守。尽管拥有牛津大学学士学位且对帆船运动情有独钟，这位老国王生前却极力展现出平民化的亲民作风。他的妻子索尼娅（Sonja）出身纺织商贾之家。他曾公开高度赞扬挪威左翼游击队在抵抗纳粹侵略中的卓著功绩，并对难民流露深切关怀。社会主义报刊《阶级斗争》（Klassekampen）曾将哈拉尔的讣告印在头版，大标题写道：“他亦是共和派的国王”。

在父亲患病期间，哈康曾摄政数年之久，然而正式执掌大宝绝非易事。他家族爆出的丑闻，让英国王室那些风流琐事显得微不足道。新王后梅特-玛丽特（Mette-Marit）在杰弗里·爱泼斯坦（Jeffrey Epstein）因诱奸未成年人定罪后，仍长期与其过从甚密（她事后为此道歉）。她与前任伴侣所生的儿子于今年6月被判强奸罪名成立。其他王室成员的表现亦称得上荒诞离奇。哈康的姐姐玛莎·路易丝公主（Princess Martha Louise）曾开办学校教授学员如何“与天使对话”。此类丑闻一度导致君主制支持率在今年2月暴跌至60%的冰点。

哈康最终可能不得不面对一个其先父刻意回避的问题。随着君主预期寿命的延长，越来越多国家的君王选择提前退位。丹麦女王玛格丽特二世（Margrethe II）于2024年退位，日本明仁天皇亦于2019年逊位。荷兰最近三任君主皆在步入暮年后交出权杖。然而哈康或许会效仿其父亲的做法退缩不前。1999年哈拉尔曾直言，退位是对传统体制的公然冒犯；除非自己彻底“老糊涂”，否则决不提前退位。

挪威议会目前正探讨制定方案：一旦君主与摄政王同时无法视事履职，允许其他高级王室成员代为行使职权。此举旨在防范类似2024年4月的行政瘫痪危机重演——当时由于哈拉尔身患重病而哈康身处海外，罢免卫生部长的决议被迫拖延搁浅。所幸新王年仅53岁，议员们尚有充裕的时间从容应对此类体制难题。■ # 欧盟东扩遭遇冰岛式暗礁

公投否决重开入欧谈判，背后是渔业利益而非地缘政治在作祟

摄影：Getty Images

2026年8月30日 | 布鲁塞尔

每年9月的第一周，是欧洲官僚们结束海滩消夏长假，重整旗鼓、准备大展拳脚应对大陆难题的时节。令人遗憾的是，大西洋中部飞来的一记冷眼打破了“秋季复工”（rentrée）的良好心境。在8月29日的全民公投中，冰岛选民以微弱差距否决了重启加入欧盟谈判的提议。尽管这次公投主要围绕本土利益关切展开，但对布鲁塞尔的决策者而言，这无疑是一记令人沮丧的“开工警钟”。对于一个习惯于探讨“别国以多快速度加入”的俱乐部而言，被当面回绝一句“不，谢谢”，着实令人尴尬。

外人或许曾预期冰岛的公投将取决于当今时代的重大地缘政治风云。战火正在欧洲大陆肆虐，而冰岛的邻邦格陵兰岛正被美国总统唐纳德·特朗普所觊觎——其派驻雷克雅未克的驻外大使曾戏言，这两座岛屿有朝一日可能成为美国的第51和第52个州。绝大多数有能力结盟自保的国家纷纷付诸行动。俄罗斯入侵乌克兰后不久，瑞典和芬兰火速加入了北约。乌克兰已正式申请加入北约和欧盟；就连加拿大——另一个被特朗普视作未来潜在“美国一州”的国家——也正竭力拉近与欧洲联盟的距离。早已是北约成员国的冰岛，看似理应顺理成章地将欧盟席位收入囊中。

然而，一切公投的根源终究在地方。占冰岛货物出口总额近40%的渔业部门，是反对加入欧盟的急先锋。早在2013年冰岛刚熬过毁灭性金融危机之后，正是渔业集团的竭力阻挠破坏了冰岛加入欧盟的初次尝试。此次他们抛出了完全一致的理据：欧盟的共同渔业政策将允许其他成员国的渔船侵入冰岛水域，并强加捕捞配额，进而重创其出口经济。反对阵营的底气还在于一种深层信念——坚信冰岛依靠自身完全能够繁荣昌盛，其与欧盟签署的既有合作协定以及北约成员国身份已筑起足够坚固的护城河。

问题还不仅出在渔业上。除了首都雷克雅未克的大多数选民支持重启谈判外，其他所有选区均呈现出压倒性的反对票势头。农场主们忧心外来竞争冲击以及政府农业补贴的缩水。企业雇主与部分工会势力同样站在了反对阵营一边。事实证明，这些担忧远比留欧阵营的游说更具说服力——后者曾极力鼓吹，一个微型岛国唯有正式融入更广阔的超国家联盟方能兴盛不衰。

布鲁塞尔方面无疑会极力淡化这一挫折。冰岛人口仅为区区40万，仅占欧盟总人口的0.1%，比当前任何一个成员国都要稀少。冰岛早已加入欧洲经济区（EEA）——相当于欧盟的外围缓冲层——同时亦是申根免签协定的成员国。此次公投原本仅是为了重启因2013年政府更迭而搁浅的技术性谈判。即便重启谈判，后续磋商也极有可能再次陷入僵局。

但对于一个长期向来以对外吸引力为傲的联盟而言，这番开脱难免显得苍白无力。目前共有10个国家（总人口达1.5亿）已正式递交了加入欧盟的申请（尽管与规模最大的土耳其之间的谈判已被冻结数年）。这些申请国全部属于贫困国家，大多政局紊乱，其中一个（乌克兰）更深陷战火泥潭。本周末的公投无情地提醒着人们：瑞士、挪威等真正富裕的潜在候选国，根本无意急于跨进欧盟的大门。

在布鲁塞尔，欧盟官员引以为傲的叙事向来是：当特朗普以吞并威胁恫吓其邻邦之际，欧盟的巨大向心力正吸引着周边国家争相加入。可惜，有人忘了把这套话术说给冰岛人听。■ # 弹丸小国黑山仍翘首以盼成为欧盟下一成员国

然而欧盟各成员国，尤其是法国，恐怕已无心扩大版图

摄影：Getty Images

2026年9月3日

西巴尔干诸国常拿自己旷日持久的“入欧马拉松”开自嘲玩笑：“我们假装推进改革，他们假装真心接纳。”然而，2022年俄罗斯入侵乌克兰为这一死气沉沉的进程注入了全新生机，黑山政府认为自己即将迎来破局的历史契机。坐拥区区63万人口的黑山，接纳它几乎不会给欧盟造成太大的财政援助负担。其农业出口体量微乎其微，不至于引发欧洲农民的恐慌，亦不会带来大规模劳工移民涌入的潜在威胁。

在今年6月于黑山召开的峰会上，欧盟领导人们争先恐后地盛赞东道国所取得的改革突破。该国政府确立的目标是在今年底前完成与欧盟所有监管标准的对齐工作，并在2028年正式加入欧盟。欧盟委员会主席乌苏拉·冯德莱恩（Ursula von der Leyen）称这一时间表“触手可及”。然而，8月29日冰岛公投否决重启入欧谈判的结果，令这股乐观情绪骤然降温。此前有人曾寄望一个富裕的北欧国家加入申请能提振扩员动能。如今，一众巴尔干申请国只能靠一己之力，去竭力说服欧洲选民：接纳贫困的东南欧国家符合欧盟自身的长远利益。

地图：经济学人

在西巴尔干一众候选国中，黑山一马当先。阿尔巴尼亚紧随其后，而科索沃以及波黑的进展则步履维艰。北马其顿遭遇了敌对邻邦保加利亚的强硬政治杯葛。作为六国中体量最大的塞尔维亚，其入欧进程在7月遭到了8个成员国的联合抵决，主要归咎于其威权主义的治理模式。8月27日，黑山总理米洛伊科·斯帕伊奇（Milojko Spajic）直言，两个邻国（显然暗指塞尔维亚与阿尔巴尼亚）出于嫉妒心理，正蓄意拖延黑山的入欧进程。

在南斯拉夫解体后，黑山曾长期由单一政党主导统治长达数十年，该党带领国家实现了独立并加入了北约。2020年，该政权被一个包含亲塞尔维亚团体的执政联盟所取代。亲欧派人士一度忧心塞尔维亚将操纵该国的外交政策。然而这一幕并未上演；两国目前的双边关系相当冷淡。加入欧盟的诱人前景，反而促使各派政治力量通力合作推进立法改革。

持怀疑态度的人士则认为这些所谓的制度变革不过是粉饰门面，并指出欧盟高层已刻意不再公开提及黑山根深蒂固且势力庞大的跨国毒品走私集团。支持欧盟扩张的欧洲人士则担忧，如果连这样一个袖珍小国都无法迈过门槛，那么乌克兰等体量更大的国家将永无入局希望。其中潜藏的最大政治风险来自法国。尽管总统埃马纽埃尔·马克龙（Emmanuel Macron）表态支持扩大欧盟，但法国宪法硬性规定，批准新成员国加入必须经由全民公投或国民议会与参议院两院五分之三的绝对多数表决通过。根据欧盟民调机构“欧洲晴雨表”（Eurobarometer）的数据，仅有35%的法国受访者支持欧盟吸纳新成员。此外，马克龙即将于明年任满离任。接替他的极可能是来自民粹右翼国民联盟的玛丽娜·勒庞（Marine Le Pen），而后者历来坚决反对任何形式的扩员。

今年6月马克龙出访黑山期间，双方签署了一揽子政治与经贸合作协议，其中一项专门旨在提升黑山在法国本土的公众形象。然而，曾任黑山欧洲事务部长的约瓦娜·马罗维奇（Jovana Marovic）对此并不抱盲目乐观。她直言，在法国国民议会内部，“根本不存在”支持批准任何潜在新成员国入欧的稳定多数票。

“尽管这番话极为不中听，”马罗维奇继续坦言，“但黑山国内的狂热氛围几乎到了盲目乐观的地步，仿佛加入欧盟早已板上钉钉。”她认为黑山尚有回旋时间在欧盟各成员国内凝聚支持共识。但就现阶段而言，许多欧洲国家对于接纳新成员的警惕防范心理，丝毫不亚于冰岛人对投身其中的抗拒排斥。■ # 美国负债惊人，但欧洲的债务问题才真正惊人

欧盟核心地带隐藏着一个“法国体量”的巨大泥潭

摄影：Peter Schrank

2026年9月3日

身为欧洲人，便意味着一边享受着精致的美食、宜步行的城市景观，一边不得不终日面对外界喋喋不休的指责——痛陈欧洲经济在诸多维度已被美国无情甩开。近年来，哀叹欧洲生产力落后的大部头权威报告，本身已然演化成一种独特的文学流派。因此，近几周传出美国联邦政府债务总额正式突破40万亿美元大关、引发债券市场对于还能维系其长期预算赤字多久的普遍质疑时，部分欧洲舆论圈内不禁泛起了一阵幸灾乐祸的快感（或者该用希腊语中对应的词汇来精准形容）。毕竟，若论及应对不断恶化的债务危机螺旋，欧洲大陆可谓经验老到。倘若攻守之势异位，轮到欧洲人对华盛顿的政客们如何治理财政指手画脚，那将是何等扬眉吐气（难道美国如今也迫切需要一份专属于自己的“德拉吉报告”吗？）。

诚然，初看起来，欧洲的公共财政状况显著优于大西洋彼岸的宿敌——这是宏观经济数据罕见地全面倒向旧大陆的高光时刻。整个欧盟成员国政府的负债总额仅为15.7万亿欧元（约合18.2万亿美元），而承载这一债务盘子的经济体量与美国相差无几（欧洲人均财富略逊一筹，但胜在人口基数更为庞大）。二十年前，两地的债务占GDP比重均在65%左右徘徊。时至今日，美国的这一比率已急剧膨胀至120%以上，国际货币基金组织（IMF）预测其到2031年将突破140%的红线。相比之下，欧洲的整体负债率“仅”为83%，未来走势基本维持平稳——尚且处于相对审慎的合理区间。然而，那些跃跃欲试想要对大洋彼岸进行经济治理说教的欧洲人，最好还是收敛锋芒。因为尽管美国的负债规模更为庞大，但欧洲面临的债务结构死结却要致命得多。美国的底气在于其掌握着全球通用结算货币且经济基本面增长强劲。而欧洲的致命短板，可以用一个词精准概括：法国。

对比欧美的债务负担，远比表面呈现的数据指标更为错综复杂。美国背负的债务固然过多，这毋庸置疑，但欧洲的债务却全部错配在最不该背负的危险环节。在应对数万亿美元联邦国债的利息偿付时，美国整体的实力远大于其各部分的机械叠加：整个联邦的财力作为坚实后盾，共同捍卫每一美元的对外信誉。相比之下，欧洲的绝大部分主权债务均由27个成员国各自独立背负，每个成员国必须独力自力更生偿还本息，否则一旦某一链条断裂，整个联盟都将面临倾覆解体的系统性风险。这一制度设计使得欧盟（或者说其最核心的欧元区架构）的抗风险极限完全取决于其最脆弱的一环——正如昔日希腊濒临主权违约时所暴露出的致命缺陷。欧盟内部诸多国家拥有极为健康的债务控制水平，尤其是德国及其北欧邻国。另一些国家（如希腊与西班牙）虽背负高额债务，但亮眼的经济增速足以安抚嗜血的“债券义警”（bond vigilantes）。然而，法国却呈现出一种极为凶险的组合：既丧失了财政施策空间（其公债总额已攀升至GDP的118%），又彻底失去了经济增长引擎（最近一个季度的GDP增长率为令人尴尬的0%）。

其直接后果是，法国十年期国债收益率如今已飙升至4.2%，创下18年以来的历史最高纪录——甚至反超了欧洲长期以来的“问题儿童”意大利。在一定程度上，这是一种全球性的宏观共振：由于投资者对通胀反弹与地缘政治风险忧心忡忡，全球各国政府都在承担更为高昂的利息支出。然而，法德两国国债收益率之间的利差——即法国为吸引投资者接盘其高风险债券所必须额外让渡的风险溢价——目前已逼近整整一个百分点。诚然，欧盟各成员国之间借贷成本的利差分化，尚未恶化到2010年代欧债危机全面爆发时的惨烈境地。这主要归功于欧洲央行（ECB），央行在此期间释放出明确干预信号：一旦某些欧元区成员国的融资成本飙升失控，央行将入场托底。但必须认清的是，欧洲央行的慷慨援助历来附加着极其苛刻的先决条件——陷入困境的当事国政府必须签署具有法律效力的整改承诺，严肃推进深层次结构性改革并彻底出清其财政烂摊子。

然而，政治现实往往是对稳健理性经济决策的最大掣肘，而欧洲眼下恰恰面临着一波凶险的政治民粹周期。2027年，法国、希腊、意大利和西班牙这四个背负沉重债务包袱的国家都将迎来全国大选。随着政客们为了讨好选民大肆挥洒竞选福利，财政纪律的防线势必将全面失守。在这场风暴中，法国无疑是最危险的震中。目前在民调中遥遥领先的极右翼民粹旗手玛丽娜·勒庞曾力主让法国脱离欧元区。近来，其所属政党的领导层更是放言，可以施压胁迫欧洲央行以某种变通手段下场兜底。而在极左翼民粹阵营，其代表人物让-吕克·梅朗雄（Jean-Luc Mélenchon）则公开叫嚣要对法国发行的一部分政府国债直接“付之一炬”。即便是那些理智温和的中间派政客，当下亦无一人敢于提议在短期内让财政赤字率重新收紧回归至欧盟条约所强制规定的3%警戒线以内。

我死之后，哪怕债浪滔天

当下断言法国将沦为“下一个希腊”，未免属于危言耸听的夸大其词——至少目前尚不至于此。然而，法国毫无节制的财政放纵却在实实在在地重塑欧洲的未来地缘经济格局。美国之所以能够长期维系巨额财政赤字，核心支撑在于其海量的短期国债全数以全球终极储备货币计价，因而在充满不确定性的动荡周期中，天然充当了全球避险资金的坚固避风港。欧洲梦寐以求想要分食这一“过度特权”（exorbitant privilege）。其尝试的重要途径之一，便是极力推动欧盟以自身名义发行由27个成员国联合担保的超级“联邦级”公共债券。该机制于2021年开始大规模破冰运作，用以向后疫情时代的经济复苏基金提供融资支持。各方此前热切期盼未来能进一步扩容此类联合举债规模以填补庞大的联合防务开支，这亦契合了法国总统埃马纽埃尔·马克龙关于构建欧洲“战略自主”的雄心蓝图。

然而，法国财政预算的持续溃败与失控，实际上已给进一步推行联合借贷的构想敲响了丧钟。法德两国国债收益率利差拉得越开，每当发行一笔联合担保债券时，从柏林暗中流向巴黎（以及更广泛地从欧盟内部财政充盈国家变相补贴给财政见底国家）的隐性财政补贴规模就越发惊人。对于从一开始就对这一机制心存芥蒂与抵触的德国而言，如今手握坚不可摧的阻击筹码：谁会心甘情愿与一个连资本市场都对其信用资质疑云密布的败家子在同一张账单上共同签字画押？长久以来，法国在财政纪律上的极度放任始终只由法国人自己买单吞下苦果。而今，其他欧洲国家也不得不被迫一同承受这场无妄之灾。■ # 英国暴力犯罪地理分布之变

在伦敦等大都市，暴力犯罪下降最为迅猛。真正令人担忧的，是那些郡县地区

插画：Gregori Saavedra

2026年9月3日

2022年1月28日，年仅18岁的乔·迪克斯（Joe Dix）倒在血泊中。他被敌对贩毒帮派的成员连刺七刀——其中一人挥舞着一把长达24英寸（约合61厘米）的长剑。一小时后，迪克斯不治身亡。“他的肠子都流了出来，”他的伴侣日后在法庭上这样回忆道。

你认为这起惨剧发生在何处？因持刀犯罪而声名狼藉的伦敦，显然是人们的第一猜想。伯明翰、曼彻斯特抑或利物浦，这些同样背负帮派暴力斑斑劣迹的城市，也是可能的答案。然而事实是，迪克斯遇害的地点是在诺里奇——一座小巧古朴的城市，就在不久前，它还被《星期日泰晤士报》评为全英最宜居城市之冠。

超过70%的英国民众坚信，全英范围内的持刀伤人事件正愈演愈烈。然而事实并非如此。作为衡量暴力犯罪最可靠的硬指标之一，英格兰因“利器袭击”入院救治的人数自2019年以来已骤降三分之一。去年，全英凶杀案发案率更是创下了50年以来的历史最低点。

社交媒体上铺天盖地的末日唱衰言论在一定程度上解释了这种认知偏差，商店盗窃与飞车抢夺手机案件的激增同样加剧了恐慌。但一个鲜为人知的深层原因在于：暴力犯罪的空间地理分布格局已发生巨变。历史上曾是暴力滋生温床的大都市核心圈，在过去十年中取得了令人惊叹的治安好转。但在广袤的郡县地区，犯罪率下降的速度则极其迟缓。在某些郡县，这一指标甚至不降反升。

以伯明翰、伦敦和曼彻斯特为例。在截至2018年3月的一年中（即英国持刀犯罪大爆发的初期），这三座城市合计占到了全英与刀具相关的凶杀及谋杀未遂案件总数的47%。而截至2026年3月的一年中，这一比重已大幅萎缩至36%。在这三座城市中，凶杀案件均已从历史峰值下降了30%以上。相比之下，萨福克郡（降幅18%）、汉普郡（14%）、诺丁汉郡（13%）等地的改善步伐要缓慢得多，而剑桥郡的案发率甚至飙升了23%。

图表：经济学人

因“利器袭击”而送医治疗的数据亦呈现出高度一致的演变规律。多数警方管区的统计数据仅更新至截至2025年3月的一年。若以两年的滚动平均值测算，伦敦的送医数量相较于2018至2019年的峰值时期已大幅锐减28%。而斯塔福德郡（微降14%）、达勒姆郡（7%）、德文郡与康沃尔郡（5%）以及赫特福德郡（5%）的降幅则要逊色得多。在萨塞克斯郡，相关数据甚至完全没有出现丝毫回落（见图表）。

尤为引人注目的是，虽然基数较低，但剑桥郡与贝德福德郡因利器致伤的送医率增速最为惊人——这两个环境幽静、绿树掩映的郊区社区，去年的送医人次分别暴增了60%和50%。米德尔斯堡是一座经济并不景气的小镇，但在历史上从未有过长期的严重街头暴力梦魇，如今其利器伤人送医数量却已彻底反超全英危机峰值时期的水平。在两名警员及另外七人接连丧生并引发日益严峻的社会动荡后，当地已沦为重大跨国跨区有组织犯罪调查的核心阵地。即便是聚焦于伦敦内部，警方记录的暴力行径增速最迅猛的区域，反而是希灵登（去年增长9%）和泰晤士河畔金斯顿（增长13%）等边缘城区，这些地带恰恰是治安向来优良的富裕郊区。总而言之，2010年代掀起的暴力犯罪浪潮，如今在传统大都市核心圈之外的土壤中展现出了最顽固的生存韧性。

这一地缘流变在很大程度上是英国毒品分销网络底层结构重塑的产物。在英国，暴力犯罪与毒品贩运犹如孪生毒瘤。这一铁律在郡县乡村与大都市中心毫无二致。例如，2021年发表的一项针对肯特郡长达六年的空中救护出警记录的研究明确揭示，“贯穿性创伤”事件的发生轨迹与毒品走私主干道高度重合。这些被称作“郡际贩毒网”（county lines）的黑金通道于2010年代末期迅猛成型，彼时暗网的兴起（更多用于毒品大宗批发而非终端零售），叠加紧缩政策下警力预算的大幅削减，导致城市核心区的毒品黑帮急剧裂变滋生。帮派之间争夺地盘的残酷内斗，直接引爆了英国2010年代末期的持刀犯罪海啸。

随着各大黑帮为避开大城市日趋饱和拥挤的市场，开始向外开辟未被开发的处女地，暴力恶疾随之被转移蔓延至广袤的郡县地区。他们构建了一套成熟的商业模式：利用心智脆弱的边缘未成年人充当“骡子”运毒——穿梭奔波于城市大帮派与乡村瘾君子之间完成终端递送。在这条黑色链条上，刀具成为了保护自身性命与毒品货物免遭敌对势力“黑吃黑”劫掠的标准配置。

值得庆幸的是，相比第一波扩张恶潮席卷之时，多数地区的暴力绝对指标已有所回落。然而，乡村地区的暴力顽疾却愈发难以根除，因为这些边缘腹地已成为贩毒黑帮最具暴利增长潜力的核心金矿。英国内政部在2020年发布的一项审查报告指出，城市毒品消费增长的见顶停滞，迫使黑帮加速转战围绕周边卫星城镇与大学校园开辟全新下沉市场。据年度安全威胁评估报告显示，被官方锁定掌握的“郡际贩毒网”数量，已由2017年的720条急剧膨胀至今日的逾6500条。毒贩们的产品线亦大肆扩容：早期的郡际网络主要分销快克可卡因与海洛因，而今大麻、高纯度可卡因粉末、氯胺酮（K粉）以及处方受管制麻醉药品已悉数被纳入其分销版图。

乡村下沉市场竞争的白热化，倒逼该灰色产业全面转向以“客户履约服务”为核心的竞争逻辑。首当其冲的便是对极速配送时效的严苛追求。在英国的诸多小城镇，瘾君子如今下单后一小时内便能送货上门——堪称安非他命界的“亚马逊Prime极速达”。面对远在马盖特（Margate）的终端客户，一家大本营盘踞在伦敦的传统跨区网络显然鞭长莫及。供应链体系被迫顺势重构。“派遣手下马仔坐火车跨城运毒的经典‘郡际网络’模式已成明日黄花，”犯罪学专家西蒙·哈丁（Simon Harding）解析道，“如今的做法是在当地直接设立前置仓集散中枢，由驻地人员就地分销，效率要高得多。”

其演变结果是，毒品帮派正日益在乡村地区实现彻底的“本土化”，他们直接盘踞在当地居民住宅内就地出货，并在当地大肆诱骗招募本土未成年人入伙。警方掌握的“郡内内生贩毒网络”（即完全在单一郡县内部闭环运作而非跨区域流窜的团伙）数量在2022年至2025年间激增了整整两倍。源自“特遣重难点区域”（即传统治安热点大城市）的外溢网络占比已从70%断崖式暴跌至50%。

然而，面对这一深刻的结构性位移，英国政府的应对举措显得迟钝脱节。今年4月初出台的《持刀犯罪行动计划》虽然在宏观辞藻上无懈可击，却沿用了针对“郡际网络”的陈旧定义——其核心承诺仍是给英国交通警察局（BTP）大举扩权，旨在截断利用铁路线从大城市向城镇贩毒的传统通道，并继续向老牌的“跨郡反毒特遣队”追加拨款。这套重拳出击所针对的，恰恰是早已在实操中发生蜕变迭代的过时商业模型。

刀刃虽藏，防线已旧

当如此庞大规模的暴力行径，是由一群被穷凶极恶的有组织犯罪集团所物色诱骗的边缘弱势青少年所犯下时，源头预防已绝非单纯的警务治安抓捕课题，而是一场不折不扣的公共卫生体系攻坚战。在昔日持刀暴力的重灾区城市，当地在长期同黑帮的周旋博弈中积累了相当丰富的防范经验，从耳聪目明的家庭成员到深谙街头规则的青年社工，早已构建起坚固的缓冲网络。然而，广袤的乡村与郡县地区却极度匮乏此类认知沉淀。许多地方仍将持刀犯罪固化为纯粹的城市特有病甚至族裔标签问题，认为单靠警力强力严打便能药到病除。“这种根深蒂固的执念极难扭转，”曾任苏格兰暴力削减中心负责人的前资深警官尼文·雷尼（Niven Rennie）坦言，“但一旦你真正参透了背后的演变机理，你就绝不会再退回到老路上。”■ # 伦敦打击持刀犯罪的治理之道初见成效

萨迪克·汗将曾经处于边缘地位的“公共卫生干预疗法”推向主流

插画：Gregori Saavedra

2026年9月3日

伦敦以诸多标签闻名于世。但在所有名声中，最不光彩的莫过于其被冠以持刀犯罪温床的恶名。去年，深陷残酷毒枭内战泥潭的墨西哥官方竟然更新了赴英旅行安全警示，郑重提醒本国公民提防英国首都街头的持刀恶性袭击。唐纳德·特朗普更曾出言讥讽，声称伦敦人整日生活在担心自己“屁股挨上一刀”的巨大恐慌中。

这番论调在历史上从未真正符合客观事实。在过去八年中的整整六年里，纽约市民遭刺杀身亡的概率事实上始终高于伦敦市民（若换算为其他暴力致死手段，纽约人的遇害概率更是至少高出十倍之多）。但不可否认的是，外界恶评并非全凭空捏造。自2017年起，英国的持刀犯罪率确曾经历过令人心惊的爆发式激增，在青少年群体中尤为严峻。在截至2018年3月的一年中，英格兰与威尔士地区共有超过280人惨死于利刃之下，远高于此前五年年均略超200人的伤亡基准线。作为这场治安危机绝对风暴眼的伦敦，在当年独占了108起持刀命案（相比之下，纽约同年死于刀具袭击的人数为62人）。

自彼时起，恶性暴力事件发生率呈现出断崖式的良性下行趋势。自峰值阶段至今，伦敦因“利器袭击”送医抢救的人数已整体锐减三分之一。持刀凶杀案件发生率直接腰斩。去年，该首都录得的青少年凶杀发案总数已降至近30年来的历史并列最低谷（比肩2012年水平）。及至2025年，身处纽约遭到利器刺死的概率已相当于伦敦的两倍整。

这当归功于萨迪克·汗（Sadiq Khan）。右翼政客向来热衷于将所谓首都街头“无法无天”的虚幻乱象归咎于这位工党籍市长。而客观事实却是，正是汗勋爵率先将一套行之有效的持刀犯罪系统性干预战略引入英格兰。这套治理机制此后已广泛推广至伦敦以外的广阔疆土。

在2017至2019年全英深陷治安恶疾而手足无措之际，汗勋爵将目光投向了北方的苏格兰。在更早的十年间，曾一度被冠以“欧洲谋杀之都”恶名的格拉斯哥，在暴力犯罪治理上交出了一份惊艳的答卷。其中最为核心的制度抓手，便是全面铺开苏格兰暴力削减中心（VRU）机制——其核心逻辑是将恶性犯罪视为一种类似传染病的“公共健康危机”，通过将警务部门、公立学校、医疗机构以及社区体育俱乐部深度整合串联，尽早筛查识别出处于犯罪边缘的高危青少年，前置性地引导其彻底脱离帮派黑洞。

在2008年至2018年间，苏格兰VRU机制的运转被公认为成功换来了凶杀案发案率大跌38%、谋杀未遂与重度故意伤害案件锐减43%的辉煌战果——其中绝大多数案件均与刀具暴力直接相关。2018年，汗勋爵顺势在伦敦拍板组建了英格兰首家VRU。时至今日，类似的暴力削减中心已在英格兰与威尔士全境扩容至20个，严密覆盖了全英80%持刀犯罪案件所集中的高危区域。

英国内政部近期公布的一项评估审查证实，自2019年体系铺开以来，英格兰设立的VRU成功促使25岁以下人群因暴力伤害送医急救的比率整体拉低了12%。相较于格拉斯哥的历史辉煌战绩，这一降幅看似温和，但这仅仅是内政部对全国大盘拉平统计的平均均值。在伦敦这类人口超密集、城市网络高度成熟的大都市圈，该机制的推行最为得心应手，暴力犯罪的降幅亦最为惊人。

诚然，评估这一成效仍需保持审慎客观。作为诱发恶性暴力犯罪的关键催化剂，酗酒行为在全英范围内整体呈降温趋势。苏格兰治理神话的铸就，在时间线上恰好与其大幅收紧酒精专卖执照法案的重磅改革高度重叠，因而绝不能被单纯归功于VRU机制单兵突进的功劳。伦敦亦是同理（在2020年年中至2024年年中期间，伦敦多达29%的夜店闭店消亡）。此外，在特定区域摸索出效能最高的落地举措亦需要长期磨合与经验沉淀。苏格兰当年的制胜杀手锏，在于直接向公立医院急诊室常态化派驻专业社工，第一时间介入干预因肢体冲突受伤入院的青少年。相比之下，英格兰与威尔士的VRU仍处于摸着石头过河的起步阶段，资金大量耗费于缺乏明确有效性证据支撑的泛化干预项目上（例如组织讲师进校园宣讲等）。

然而，贯穿其始终的底层制度逻辑却坚不可摧。慈善智库青年发展基金会（Youth Endowment Fund）于2023年发布的一项实证研究揭示：童年时期若遭遇过两起或以上的恶性创伤经历，其日后滑入黑帮深渊的概率将直接激增61%。在当下的英国，恶性街头暴力在本质上高度受控于将边缘脆弱青少年当做运毒工具的跨区域有组织犯罪网络。这与2000年代初期的格拉斯哥有着本质不同——当年的街头地盘流氓更热衷于发泄性的互殴械斗而非成体系的毒品分销走私。但无论犯罪生态如何演变，VRU的精准介入都能釜底抽薪，从源头上大幅扼杀并斩断黑帮用于压榨剥削的廉价底层劳动力供应池。

在犯罪学专家伊恩·布伦南（Iain Brennan）看来，苏格兰当年最可贵的精髓，不仅在于引进了这套机制工具，更在于从根本上重塑了全社会所有参与者的价值认同与精神理念。“其灵魂在于彻底转变对暴力防治的哲学认知，扭转整个社会的应对范式才是至关重要的一环。”汗勋爵时常因推行看似软弱温和的治理套路而遭到政敌攻讦讥讽——他总在不知疲倦地强调开办青年活动中心以及竭力保住问题少年学籍的重大意义。然而，正是这套看似春风化雨的“温和拳法”，却在硬桥硬马地击碎顽疾、打出实打实的硬核战果。■ # 伦敦的手机飞车抢劫潮已越过峰值了吗？

曙光初现的积极迹象

插画：Gregori Saavedra

2026年9月3日

如今的伦敦较以往已安全得多。然而，一种特殊的犯罪门类近年来却呈现出野蛮疯涨的态势：侵财性盗窃。自2020年以来，无论是入店行窃还是“针对人身的盗窃”（其中绝大多数为街头飞车抢夺手机），案发总量几乎翻了一番。在截至2026年7月的年度周期内，入店盗窃发生率再度刷新历史天花板。然而，衡量伦敦个人随身财物被盗情况的12个月滚动累计数据，已从2025年1月的历史最高峰大幅回落了27%（见图表）。

图表：经济学人

这一改善在很大程度上归功于警方缉捕战术的迭代升级。手机飞车党普遍专精于“突袭式掠夺”——骑乘大功率电动自行车无声无息地逼近低头看手机的目标行人，得手后瞬间狂飙遁形。对于警方而言，围捕此类流窜劫匪曾是一场注定败北的猫鼠游戏。警方的硬性指标是在接到紧急报警后15分钟内赶赴现场。然而对于一辆时速高达50英里（约合80公里）的电单车而言，这点时间足够其穿街走巷、借道逼仄小巷或在滚滚车流中辗转腾挪，轻而易举地将笨重庞大的传统警用车辆甩得无影无踪。

警方战法被迫顺势求变。伦敦大都会警察局一方面大举组建专属的追击电单车车队，力求以快制快追捕嫌犯；另一方面加大查扣力度，对犯罪分子用于作案的非法改装电单车实施地毯式的大规模收缴。自去年10月起，警方更是祭出高空侦察无人机战术，在安全距离外实施超视距追踪，静待飞车贼彻底放松戒备后瓮中捉鳖。配合在飞车抢劫最高发地段（诸如威斯敏斯特区等游客密集的黄金地带）常态化派驻定点巡逻警力，这套打法组合拳已显现出立竿见影的遏制成效。

此外，一线办案警官亦已彻底摸透了手机飞车劫掠背后的灰色产业链运行机理。看似零散孤立的街头偷摸扒窃，其实质高度依附于一条由跨国专业销赃网络主导的出口链条，将海量被盗手机走私至海外。因此，在链条顶端收网逮捕极少数核心操盘手，便能取得四两拨千斤的打击成效。去年10月，大都会警察局一举端掉了18名涉案关键嫌犯，这伙人涉嫌秘密将多达4万部被盗手机走私贩卖至中国市场（这一数字竟占到了当时伦敦被盗手机总数的大半壁江山）。

几乎每一个生活在伦敦的人，身边都有亲友遭遇过手机被抢的倒霉经历。这在极大程度上解释了为何该首都乃至全英公众的社会不安全感持续居高不下。入店行窃目前依然是屡禁不绝的社会顽疾。但至少在应对飞车掠夺手机这一特定毒瘤上，伦敦似乎已找准了对症下药的破局正道。■ # 安迪·伯纳姆能否将连胜势头延展至伦敦？

一场关键的议会补选，将全面检验这位新任首相究竟是货真价实的全民领袖，还是仅昙花一现的“北方现象”

摄影：AP

2026年9月3日

2026年注定被载入英国当代政治史的“补选之年”。今年2月，心怀怨怼的曼彻斯特选民通过选出一名绿党议员，在基尔·斯塔默（Keir Starmer）爵士的首相政治棺木上钉下了致命一击。到了6月，来自英格兰北部的梅克菲尔德（Makerfield）选区以压倒性选票将安迪·伯纳姆（Andy Burnham）送入议会，并赋予其全面接替斯塔默出任首相的民意授权。7月，在伯纳姆入主唐宁街10号正式执掌乾坤后，大曼彻斯特地区顺理成章选出了一名工党籍新市长接替其空缺，进一步巩固了这位新首相无可匹敌的竞选号召力。

如今，轮到伦敦选民登场亮牌。9月1日，斯塔默爵士正式宣布辞去议员职务。此举旋即引爆了其原先所在的伦敦霍本和圣潘克拉斯（Holborn and St Pancras）选区补选，该选区版图横跨游客如织的考文特花园、熙攘喧嚣的卡姆登镇以及豪宅林立的樱草花山。这场即将拉开帷幕的选举角逐，是伯纳姆向全英证明自己绝非仅仅局限于英格兰西北老家本土、亦能跨越南北鸿沟并将胜利果实延展至南方腹地的一场绝佳大考，他必须借此在工党的传统城市腹地坚决狙击绿党的凶猛渗透。

在寻常时期，霍本和圣潘克拉斯选区向来是工党稳如磐石的铁票仓。该区域人口结构年轻、种族高度多元化且受教育程度极高，自1964年以来便牢牢被工党攥在手中。然而在2017年至2024年间，工党在该选区的得票率却经历了一场惨烈的暴跌，由原先的70%断崖式滑落至49%。左翼亲巴勒斯坦活动家安德鲁·费恩斯坦（Andrew Feinstein）作为独立候选人异军突起，精准吃下了当地庞大的青年学生与穆斯林群体选票，一人斩获了19%的选票。与此同时，绿党亦强势拿下了10%的份额。绿党言辞犀利的党魁扎克·波兰斯基（Zack Polanski）已正式对外宣布，将御驾亲征出战此次议会补选。

一个高度整合集结的左翼激进阵营向工党发起正面对决，势必令工党感到头疼。在今年5月举行的伦敦地方选举中，绿党在当地横扫了大约29%的选票，而费恩斯坦麾下的参选团队亦拿下了6%的版图。即便如此，在斯塔默执政声望跌至历史绝对谷底的至暗时刻，工党在当季的地方选举中依旧咬牙保住了38%的得票率底线。此后民调走势显示，伯纳姆凭借个人魅力，已从激进阵营手中成功挽回了相当大一部分心灰意冷的偏左翼倒戈选民。

任何补选向来充满了不可预知的变数，尤其是波兰斯基的亲自出马势必在年轻选民中掀起巨大的狂热声浪。但抛开暗藏的滑铁卢风险不论，霍本和圣潘克拉斯选区的补选归根结底是伯纳姆展示统御力的绝佳历史跳板。工党若能在此斩获一场酣畅淋漓的决胜完胜，不仅将重挫当下已现颓势的绿党锋芒，彻底粉碎“新首相魅力不出英格兰北方”的地域偏见流言，更能无限延展其上任之初高光耀眼的执政蜜月期。这场政治蜜月，显然还远未走到曲终人散的终局。■ # 全球女足第一人为何屈尊加盟伦敦无名小卒？

背后的核心关口在于：一支女足球队能否彻底脱离男足俱乐部的附庸庇护而独立走向繁荣？

摄影：Alamy

2026年9月3日

阿莱克西娅·普特利亚斯（Alexia Putellas）被球迷尊称为“女王”（la reina）。在效力巴塞罗那女足长达14年的传奇岁月里，这位西班牙女足领袖先后四度捧起女足欧冠奖杯，十次加冕西班牙国内联赛桂冠。她率领西班牙国家队在2023年傲然问鼎女足世界杯，并两度荣膺代表世界女子足坛最高个人荣誉的金球奖（Ballon d’Or Féminin）。然而在今年夏天，她却出人意料地作别豪门巴塞罗那，转投主场坐落于布罗姆利、成立仅七年且仅拥有一座逼仄小球场的伦敦城雌狮队（London City Lionesses，简称LCL）。这究竟是何缘由？

LCL最初的前身是米尔沃尔俱乐部麾下的女足分支机构——而该男足俱乐部最广为人知的标签，却是其极端球迷对暴力斗殴的特殊嗜好。2019年，米尔沃尔雌狮队毅然决然选择与母体剥离分家走向独立，并改换门庭更名为现名。四年后，该俱乐部被美国女企业家米歇尔·康（Michele Kang）出资收购，后者名下同时兼并拥有美国和法国的顶级职业女足球队。自去年成功冲入英格兰女子超级联赛（WSL，英格兰女子最高级别职业联赛）以来，LCL在转会市场上掀起了一场瞠目结舌的金元狂潮。除了重磅招揽普特利亚斯入局，该队还先后砸下重金引进了英格兰前主力门神玛丽·厄普斯（Mary Earps），以及多位来自澳大利亚和法国的国际巨星。

坐拥腰缠万贯的海外资本大鳄不惜血本砸钱烧金，在英格兰男子职业足坛早已是一出屡见不鲜的标准戏码。然而，在曼城或切尔西凭借海量资本输血换来战绩荣耀的背面，亦有莱斯特城或朴茨茅斯等因无节制挥霍引火烧身、最终在财政破产边缘苦苦挣扎的悲剧前车之鉴。WSL官方正竭力防范此类暴雷风险在女足赛场重演。在定于9月4日正式开打的2026-27赛季中，所有参赛俱乐部必须严格在全新出台的“财政支出紧箍咒”规则下运转。各队在引援转会费与球员工资上的总开支，严禁突破俱乐部总营收的80%红线，外加俱乐部注资人自掏腰包的补贴额度（补贴上限为营收总额的25%或最高不超过400万英镑，约合550万美元，两者以高者为准）。一旦有俱乐部越过雷池违规，将面临严厉的经济重罚乃至直接扣除联赛积分的毁灭性制裁。

这套全新出台的刚性法案旨在倒逼女足俱乐部追求长期健康的财务自律与造血平衡，但客观上也极易固化联赛既有传统豪门垄断优势的马太效应。以阿森纳和切尔西两强为例，两家巨擘在过去九个赛季中合力霸占了八次WSL冠军宝座，在各自最新公布的财报周期中，两家俱乐部的年营收均已高达2500万欧元（约合2100万英镑或2900万美元）。面对中小球会，他们在引援转会和球员工资预算上拥有压倒性的碾压优势。

米歇尔·康正在进行一场豪赌——押注其俱乐部的商业开发运营团队拥有能在短期内让造血营收呈爆发式几何级翻倍的能力，以此全面抹平引入天价新援所背负的沉重开支包袱。在2024-25财年，LCL的全年营业总收入仅略微超过区区90万英镑。而据媒体披露，仅普特利亚斯一人的一年薪水便已逼近这一总额。尽管伴随转播版权水涨船高（天主教电视台Sky与BBC于2025-26赛季起签署了天价女足转播新约），该俱乐部的营收会有显著抬升，但LCL的运营成本开支同样呈几何级狂飙。财报显示，该俱乐部在2024-25赛季直接爆出了高达1060万英镑的巨额账面亏损。若想在新赛季中完全合规避开WSL的严苛惩罚红线，该俱乐部的商业营收指标或许需要疯狂暴增整整15倍之巨。

值得玩味的是，在这座由14支球队构成的顶级联赛殿堂中，LCL是唯一一家完全未与任何大型男子职业豪门实施行政与资本捆绑的孤勇独立球队。朴茨茅斯大学足球金融学专家克里斯蒂娜·菲利普（Christina Philippou）研判认为，这一看似孤立无援的底色反而可能是其最大的破局杀手锏：“相比于那些传统附属球队，伦敦城雌狮队能更加肆无忌惮地精准锚定并深度赋能纯女性消费品牌，而附属球队受制于男足母体传统标签的绑架与利益拉扯，往往根本放不开手脚。”她同时补充指出，米歇尔·康名下同时还牢固掌控着美国的华盛顿精神队（Washington Spirit）与法国的里昂女足（OL Lyonnes），这种多俱乐部联合运营矩阵赋予了其在全球赞助招商时，能够打包跨越三大主流发达国家市场的稀缺协同优势。

然而，形单影只的独立发展模式，同样令LCL在至关重要的硬件基础设施护城河上暴露出肉眼可见的致命硬伤。阿森纳女足目前已将全部联赛主场全面搬迁至宏伟的酋长球场（Emirates Stadium），其在2025-26赛季的主场平均上座人数已豪迈突破3.3万人大关。而相比之下，LCL眼下只能寄人篱下，通过租借英格兰第五级别半职业男足球队布罗姆利（Bromley）那座寒酸狭小的球场作为临时栖身之所。在上一赛季中，其场均到场观赛的拥趸仅仅勉强维持在区区3200人上下。

知名足球大数据咨询机构Analytics FC首席执行官亚历克斯·斯图尔特（Alex Stewart）直言不讳地指出：“若想强行打破WSL既有的豪门寡头割据格局，唯一的破局之道就是用排山倒海的资本硬生生砸开一条血路。”康女士无疑将这一铁血金元信条践行到了极致。从理论推演，普特利亚斯这枚超级重磅棋子的空降，极有可能成为彻底将LCL淬炼为女子职业足坛常青竞争力量的催化剂。在竞技绿茵场上，她的无双统治力完全有希望单核驱动这支青年军杀入欧冠联赛席位，进而为俱乐部撬开全新的天价赛事奖金金矿；而在绿茵场外，她无可比拟的全球顶流个人IP声量，完全具备招揽引爆天量商业跨界赞助的破圈能量，助推俱乐部在财务合规审计大限到来前惊险通关。但这毫无疑问是一场刀尖舔血的凶险豪赌。■ # 英国政府借助网红大V传递施政声音

然而这绝非清晰政策制定的替代品

摄影：Getty Images

2026年9月3日

安迪·伯纳姆（Andy Burnham）的早年岁月是伴随着浓郁的报纸油墨芳香浸润长大的：孩提时代他曾兼职走街串巷送报，大学毕业后的首份生计便是在《米德尔顿卫报》（Middleton Guardian）担任（无薪）见习记者。然而，自问鼎英国首相大位以来，他却未曾对传统的“纸媒行当”展露出半点香火温情。相较于传统严肃报刊，他接受各类播客节目专访的频率要高得多；比起正襟危坐的传统新闻发布会，他更是频繁在TikTok上大搞花式网络作秀。

纵观历史，每一位卓有成效的政治领袖皆是其所处时代媒介工具的弄潮儿。19世纪中叶主政的帕麦斯顿勋爵（Lord Palmerston）曾与《晨邮报》（Morning Post）维持着如胶似漆的亲密同盟，甚至直接口述授意该报的社论文章；1960年代的哈罗德·威尔逊（Harold Wilson）率先将政府公关带入电信革命新纪元，在唐宁街官邸内全面架设用于滚动播报国际即时要闻的电传打字机；玛格丽特·撒切尔（Margaret Thatcher）曾欣然接受先锋叛逆的流行文化杂志《Smash Hits》的深度专访；而托尼·布莱尔（Tony Blair）爵士更是直接将“舆论操盘手”（spin doctors）安插进国家决策中枢，以从容应对24小时全天候滚动新闻时代所掀起的信息洪流。

如此观之，伯纳姆倾注巨大心血打造极具视听张力的全套数字化公关产品线，便不足为奇了。自入主唐宁街10号以来，他平均每天在TikTok平台上发布超过一条原创短视频。其发布的网络视频内容五花八门，从一本正经地为英式传统酒馆经典小食逐一打分排名，到毫无保留的“有问必答”互动问答。即便是最棘手的坏消息，亦被其全面搬上社交网络平台分发宣泄。伯纳姆曾直接在Facebook发布长帖，坦承旨在通过对部分轻罪罪犯实施提前释放以缓解全英监狱床位危机这一争议方案在落地中暴露出的种种严重缺陷。海量网民跟帖留言，高度盛赞这位新任首相勇于向公众坦诚布公的政治担当。

在这张公关大网中，更有各大当红网络红人的身影。伯纳姆曾与知名油管（YouTuber）博主埃利斯·普拉滕（Ellis Platten）大聊复古足球球衣文化；他曾同在TikTok上因拍摄公交车日常视频而狂揽数百万播放量、坐拥“公交大妈”绰号的贝米·奥罗久贡（Bemi Orojuogun，见图）并肩合影互动；他甚至还与手握科学硕士学位、专精于在Instagram平台上发布趣味科普视频的网红“大曼尼”（Big Manny）挽起袖子联手拼装了一套物理电路。

里希·苏纳克（Rishi Sunak）是英国历史上首位正式向各路网络博主敞开唐宁街10号大门的首相，旨在借此强行穿透并触达日益远离传统主流媒介阵地的年轻选民（据全英通信监管机构Ofcom权威监测数据显示，35岁以下年轻受众平均每天收看传统广播电视节目的时长已不足1小时，相比之下65岁及以上的老年群体则高达5小时以上）。基尔·斯塔默爵士执政期间更是变本加厉，不仅创设了专供网红参与的独立新闻发布会，更在出访海外时直接随团携带一批“内容创作者”伴行。

如今在英国白厅（Whitehall，代指英国政府）官僚机构内部，将网红大V摆在与传统殿堂级媒体平起平坐位置的机制，已被彻底予以制度化确立。在《经济学人》获得的一份由政府内部起草的政策指导文件中，赫然白纸黑字告诫全体公务员：“面对来自政府冷冰冰的行政说教，公众往往会产生本能的抵触心理”，但文件随即笔锋一转：“倘若这些指导方案或行动倡议是经由一位具有亲和力、让公众产生共鸣的网红博主之口道出，其观感便不再像是居高临下的行政命令，而更像是一次心有灵犀的共同探索或良心安利。”该份解密文件同时披露，作为深度参与英国政府定制宣传战役的酬劳回报，网红每发布一条推广贴文便能稳稳入账“数千英镑”的巨额报酬。“政府有法定义务在公众聚集流连的核心场域与他们建立链接，”一名官方发言人对此坚决辩护道。

曾参与真人秀《恋岛》（Love Island）录制、如今转型为网络安全公益行动家并多次出入唐宁街的乔治娅·哈里森（Georgia Harrison）指出，现代政治博弈与社交网络底层文化本就具有极高的天然契合度。“其精髓都在于制造极具冲突感的戏剧张力，”她直言。个人理财领域知名网红阿比盖尔·福斯特（Abigail Foster）则犀利指出，网络意见领袖具备传统报章无论如何也无法企及的垂直渗透穿透力：“我们日夜活跃在下一代年轻人扎堆的阵地，我们在面容神态和谈吐口吻上，与他们本就是同一个模子刻出来的同路人。”

然而，随波逐流盲目迎合社交网络瞬息万变的风口，绝非维系政治声望的长生灵药。苏纳克在2024年被全英选民毫不留情地轰下台，而斯塔默此前处心积虑打造的花哨短视频，亦未能成功扭转其木讷僵硬的官僚刻板形象。“基尔骨子里的气质大概更贴近严肃长文平台Substack，而非短视频群魔乱舞的TikTok，”曾任斯塔默唐宁街公关总监、现任彭塔集团（Penta Group）高级合伙人的詹姆斯·莱昂斯（James Lyons）坦承。福斯特亦透露，各政府部门此前向其抛出的商业联名合作邀约过于频繁密集，以至于她后期不得不开始主动婉言谢绝——她警示称，网红受邀在唐宁街10号深宅大院内掌镜自拍所带来的那点新鲜猎奇感，在受众眼中早已消耗殆尽。

不可否认，英国各大传统报章至今仍保有不可小觑的舆论定盘星力量，其议程设置往往直接决定了老牌国家广播巨头BBC的选题取向，并在网络舆论场掀起二次发酵的惊涛骇浪。但毫无疑问的是，任何一套对社交媒体熟视无睹的政治公关体系，在当今时代都注定走向彻底的失败破产。曾参演BBC大热智斗真人秀《背叛者》（The Traitors）、如今深耕国际政治科普短视频的原外交官亚历山大·德拉戈内蒂（Alexander Dragonetti）指出，互联网催生了“超精准的圈层定制化”模式；当权者必须全方位深耕细分垂直的小众受众群体，方能确保其政治声音能够穿透各温室气泡、最终覆盖全社会绝大多数选民。

媒介与内核

在网络政治表达这一竞技场上，英国政客同大洋彼岸的美国同行相比依然存在肉眼可见的差距——在美国，两党候选人躬身落座播客直播间连续畅谈数小时早已成为稀松平常的标准操作。但迄今为止，伯纳姆凭借其在镜头前展现出的从容自若、略带接地气市井平民范（blokey）的立体人设，赢得了其前任梦寐以求的赞誉——尽管其麾下公关团队的绝大多数幕僚骨干，恰恰正是当年为斯塔默效力的同一批原班人马。

归根结底，政策内核与传递信息的领袖特质本身，其重要性永远凌驾于信息传递的媒介载体之上。在英国现代政治史上，公认对传统媒体舆论操盘得最为炉火纯青的两位首相——撒切尔夫人与布莱尔爵士，恰恰也是对自己究竟意图达成何种施政图景、以及该通过何种路径攻坚克难拥有最清晰战略定力的两位政治家，这绝非历史巧合。沉溺于媒体曝光细枝末节的领袖，往往极易在喧嚣中迷失本心、丧失战略航向。众多网红博主亦反复极力澄清，自己绝非挽救无能政府于水火之中的救命万灵丹。正如哈里森所言：“治国政策永远必须高居首位，舆论公关只能退居其后，绝无本末倒置的道理。”■

ch027.xhtml

法国正在对英国展开“糖衣炮弹”攻势

这一次，英国似乎也罕见地积极投桃报李

图片来源：盖蒂图片社 (Getty Images)

2026年9月1日 | 伦敦与巴黎

对英国人而言，很少有比遭遇法国人的“糖衣炮弹”更令人不知所措的经历了。英吉利海峡对岸的这些邻居，在英国人眼中通常是难缠的对手、令人抓狂的伙伴，而且衣着品味还招人嫉妒地考究。然而本周，法国却使出浑身解数来向英国示好。

9月2日，埃马纽埃尔·马克龙总统与查尔斯三世国王在大英博物馆一场备受瞩目的活动中会面，共同为贝叶挂毯展揭幕——该展览注定成为本季伦敦的重磅盛事。这件描绘了1066年诺曼征服英格兰历史的11世纪杰作，是有史以来首次离开法国本土展出。马克龙去年曾打趣道，借出这件国宝的谈判耗时“恐怕比脱欧谈判还要长”；他已将此次出借视为双边外交的核心个人手笔。在2025年宣布办展计划后，他谈及英法关系时表示：“只要我们坐到同一张桌子旁，一切皆有可能。”

法国为何对英国如此示好？作为历史上的宿敌，两国近年来摩擦不断：从英国脱欧、潜艇军售，到偷渡小艇等纷争不一而足。在被问及马克龙究竟是敌是友时，曾在英国首相之位上短暂停留的莉兹·特拉斯曾回答：“尚无定论。”如今，英国似乎也渴望投桃报李。在基尔·斯塔默爵士任内，马克龙获得了国事访问的高规格接待、温莎城堡的国宴礼遇，以及共同组建支援乌克兰“自愿联盟”的伙伴。9月3日，马克龙按计划成为斯塔默的继任者安迪·伯纳姆入主唐宁街后接待的首位外国领导人。外界预计两人将讨论乌克兰局势、防务安全以及英国与欧盟的关系。

解开这轮“新修好”局面的线索，或许并不在大英博物馆，而是一部精湛的全新法国两集人物传记片《戴高乐战役》（La Bataille de Gaulle）。该片由前外交官安托南·博德里执导，罕见的是，作为一部法国电影，它却改编自英国历史学家的传记著作《法兰西的某种理念：夏尔·戴高乐的一生》（A Certain Idea of France: The Life of Charles de Gaulle）。这本厚达928页的巨作由伦敦玛丽女王大学历史学家朱利安·杰克逊所著，凝聚了他多年严谨细致的档案研究心血。

对英国人的某种理念 © 盖蒂图片社 (Getty Images)

这部银幕巨作由西蒙·阿布卡瑞安领衔主演戴高乐，西蒙·拉塞尔·比尔爵士则惟妙惟肖地塑造了一个脾气乖张的温斯顿·丘吉尔。影片自然是一部关于勇气与道德操守的动人史诗——这也是围绕这位抵抗运动领袖与现代法国缔造者延续至今的传奇光环。但或许正是得益于外国传记作家的视角，影片同样客观平实地审视了这位“戴高乐将军”兼具远见卓识的伟大与身为人性的脆弱。

尤其能引起英国观众共鸣的是，影片第一部《铁血时代》（L’âge de fer）着重刻画了1940年法国沦陷于纳粹德国后，戴高乐流亡伦敦北部汉普斯特德的岁月。它毫不掩饰这位心高气傲的将军与丘吉尔之间电闪雷鸣的关系。陆军大臣安东尼·艾登曾询问首相：“押注戴高乐似乎有点孤注一掷，你不觉得吗？”丘吉尔反驳道：“这可不是我经历的第一场战争，”随后冷静补充道：“要是他失败了，我们就换掉他。”这位将军与英国首相互相发泄了大量冷嘲热讽，但也曾在彼此身上看到过理解、敬佩，以及一丝隐约的惺惺相惜。

毕竟，伦敦为戴高乐提供了避难所、办公室（卡尔顿花园4号）、话筒（英国广播公司）、支持（大多数时候）以及合法性（尽管当时美国人对此百般抵触）。法国人从未遗忘这份恩情。2020年，马克龙“破例破规格”授予伦敦市法国最高荣誉——法国荣誉军团大十字勋章，以表达“永恒的感激之情”。他表示，伦敦是“在一切看似无望时，希望的最后堡垒”。

马克龙似乎在传达这样一个信息：抛开脱欧恩怨，抛开唐纳德·特朗普的变数，或者正因这两者并存，英法两国现在比以往任何时候都更需要像当年那样紧密团结。这部在法国吸引了逾500万观众走入影院的戴高乐传记片，不仅唤起了对那个英雄时代的怀旧之情，更切中了当下有关国家主权、外部依附以及民主价值观的深层忧虑。

秉持倔强独立意识的戴高乐，曾因罗斯福战后对法国实施美军军管的秘密计划而勃然大怒。影片真实再现了丘吉尔粗暴告知戴高乐的瞬间：英国永远会选择站在美国一边，而非欧洲。然而，戴高乐所代表的战略自主教条在当下重获强烈共鸣——如今特朗普治下的美国正扬言任由欧洲自生自灭。尽管面临右翼民粹主义政党“改革英国党”的步步紧逼，伯纳姆仍公开表态英国需要在走近欧盟方面“更加大胆”。特朗普的举动客观上已经推升了英法的走近步伐：去年两国强化了核威慑合作，并共同领导着援乌的“自愿联盟”。

这部影片还传递了另一重深意：民族主义的黑暗浪潮并非不可战胜。如今民粹右翼势力在两国皆呈抬头之势，玛丽娜·勒庞是赢得明年法国总统大选的头号热门。9月4日，她的得力干将若尔当·巴尔德拉作为“改革英国党”党首奈杰尔·法拉奇的特邀嘉宾，出席了该党在伯明翰举行的大会。马克龙时常告诫法国民众切勿屈服于“失败主义心态”——这是在致敬被盖世太保枪杀的抵抗运动战士、历史学家马克·布洛赫关于1940年法国沦陷的经典著作。历史的对照或许并不完全对等，但这毫无疑问是英法必须携手并肩应对的又一场战役。■

ch028.xhtml

工作如何让家庭更平等

对金钱的追求，往往能催生诸多善果

图片来源：路透社 (Reuters)

2026年9月3日 | 勒克瑙、利雅得与斯德哥尔摩

十年前，沙特女性除极少数岗位外几乎被完全排斥在职场之外。随后，沙特实际统治者、王储穆罕默德·本·萨勒曼废除了绝大多数限制——甚至赋予了沙特女性驾驶权。他的目标是让王国减少对石油红利的依赖，转向依靠全社会的勤劳付出。让过去被困守闺房的半数人口走入职场，显然是最合逻辑的起点。

这引发了翻天覆地的转变。利雅得的年轻女性哈姆达听到这个消息时欣喜若狂，随即告诉丈夫自己想找一份工作。丈夫当即断然拒绝：如果她外出工作，就可能接触到其他男人，从而让家族蒙羞。

她与丈夫争执了一整年。最终她赢了，原因有二：其一，多一份收入确实解了燃眉之急。“我告诉他，我们得付各种账单，”她回忆道。其二，她丈夫的姐妹们全都陆续找到了工作。突然之间，丈夫再也无法在家庭餐桌上声称“只有不守妇道的女人才会工作”了。于是，哈姆达在一家商场找到了卖睡衣的工作。她说，自己的生活比以前好得多；未来有一天，她打算买一辆属于自己的车，并且自己创业开店。

在世界许多地方，女性面临的最大抗争并非争取同工同酬，而是争取走出家门工作的基本权利。斯坦福大学的爱丽丝·埃文斯指出，在中东、北非和南亚的广袤地区，“男性的荣誉在很大程度上维系在女性的深闺幽闭之上。”父亲、兄弟和丈夫们坚信，女性涉足外部世界会颠覆自然秩序，并让她们的贞洁陷入危险。

此类陈规陋习让整个地区陷入了本可避免的贫困之中。在南亚，仅有三分之一的女性参与劳动力市场；在中东和北非，这一比例甚至仅为五分之一（相比之下，北美有七成女性工作）。世界银行测算，消除女性就业的障碍可使许多国家的人均收入增加五分之一——这带来的经济效益甚至超过了避免一场内战。

益处不仅于此。正如哈姆达的经历所展现的，当曾经被迫留在家中的女性开始获得收入时，家庭内部的权力天平便发生了倾斜。工作让家庭变得更加平等。

《经济学人》通过数据分析探讨了性别歧视观念与女性劳动力供给之间的联系。我们采用了得克萨斯农工大学的瓦莱丽·哈德森以及杨百翰大学的唐娜·李·鲍文和佩佩图阿·琳·尼尔森共同构建的“父系/兄弟同盟综合征”指标。该指标涵盖了极端的男性沙文主义表现：包括家庭法与财产权对女性的不平等待遇、女孩早婚、一夫多妻、重男轻女以及对针对女性暴力的容忍度。我们将该指标与调查中低收入国家女性的“人口与健康调查”（DHS）数据进行了比对分析。

图表来源：《经济学人》

我们发现了极强的相关性（见图表）。在性别歧视最严重的21个国家中，仅有30%的女性参加工作，且无论家庭或国家多么贫困，这一比例都保持在极低水平。随着性别偏见逐渐消退，参与工作的女性越来越多，尤其是在最需要收入补贴家用的贫困地区。在性别歧视处于中等水平的贫困国家，最贫困家庭的女性就业率上升至55%，而在性别歧视最轻微的地区，这一比例则达到80%。

接着，我们研究了赚取薪资是否改善了女性的家庭地位。我们采用了由巴西佩洛塔斯联邦大学的费尔南达·埃弗林和阿卢西奥·巴罗斯等人开发的“基于调查的女性赋权指数”（SWPER）。该指数利用DHS数据来评估女性在家庭中的决策权大小：她们能否自主决定家庭预算的花销？她们外出探望亲友是否无需事先征求许可？简而言之，她们的意见是否有分量？

这些分析结果同样令人瞩目（见图表）。即便在控制了国家的性别歧视程度和家庭财富水平之后，有工作的女性在家庭决策中始终表现出更高的发言权。

背后的原因多种多样。在要求女性顺从的社会中，外出工作可以让她们学会更加自信与果敢——哈姆达日益增强的底气便是一个缩影。此外，当女性能够贡献家庭收入的很大一部分时，她们在开支支配上自然更容易拥有发言权。最后，经济独立的女性拥有离开一段不美满婚姻的底气，或者能构成切实的可信威胁。因此，丈夫们必须对妻子更好，否则就将面临打光棍的风险。“金钱就是权力，”库尔德裔瑞典作家萨基内·马东指出，“女性一旦获得更多独立，她们就能主动从糟糕的婚姻中脱身。”

打破传统禁锢

在一些地方，阻碍女性工作的壁垒来自体制与法律。根据世界银行的《女性、商业与法律》指数，约有5.4亿女性生活在法律平等依然“遥不可及”的国家。耶鲁大学的佩妮·戈德堡发现，一个国家在该指数上的得分毫不意外地与其职场女性的发展表现“高度正相关”。

因此，正如沙特所证明的，修改法律行之有效。但这仅仅是第一步。世界银行的凯瑟琳·比格尔指出，社会习俗往往是更为顽固的挑战。诸如“谁负责早上捣碎面糊，谁确保下午五点半晚饭准时上桌；这类琐碎分工并不属于政府传统政策能轻易干预的范畴。”

在一个保守的街区或村落里，第一批向旧观念说不的女性，往往会面临巨大的噤声与顺从压力。以印度北部北方邦的社区工作者桑迪亚为例，她成长于一个“男人绝不下厨”的印度教乡村。她在15岁时便嫁作人妇，丈夫禁止她在没有男性亲属陪同的情况下前往任何地方，一旦她顶嘴便拳脚相加。

六年之后，她带着儿子离开丈夫，搬回父母家生活。为了自食其力抚养儿子，她开始打工，起初做护士。然而恶毒的流言蜚语随即传开：她为什么下班这么晚？为什么穿得漂漂亮亮？难道她是妓女吗？流言未能扑灭她的上进心，但却足以让当地其他女性在寻求自立前望而却步。

图表来源：《经济学人》

在印度农村乃至许多地方，流言蜚语是维护旧秩序的利器。男性通常能通过进城打工获得更高工资，但任何尝试效仿的女性都可能被扣上性工作者的帽子，导致她们难以成婚。因此，印度的城乡流动务工者中几乎全为男性。这使得流动工人的聚集地对女性而言危机四伏，也给了男性反对女儿离村的借口——声称这是“为了保护她”。如果这意味着她必须留下来洗衣做饭，他们更觉得两全其美。

在性别偏见根深蒂固的国家，推动女性就业的技术官僚手段往往收效甚微。托育补贴或许在北欧女性身上能创造奇迹，但在埃及，当一项实验向母亲们提供能够覆盖托儿所部分或全部费用的代金券时，仅有11%的人领取使用。大男子主义的丈夫往往彻底压制了廉价托儿所带来的便利。

然而，观念终究会变。世界银行丹尼尔·哈利姆、迈克尔·奥沙利文与阿比拉莎·萨哈伊的一项研究发现，金钱、科技、社会风潮的传染效应以及地理迁移，都能促成改变。

首先是金钱。当女性外出工作时，男性亲属起初可能觉得有辱门楣。然而，即便是最顽固保守的男人也无法抗拒多一份进账的诱惑。埃文斯博士将此称为“荣誉与收入的权衡取舍”。

不同社会处理这种取舍的方式大相径庭。中国果断选择了增加收入。（新民主主义革命与社会主义改造时期通过推动妇女进入生产领域打破了旧式父权传统；随后的改革开放更是让那些将女儿送进工厂打工的家庭获得了丰厚的经济回报。）这或许解释了为什么尽管中国的适龄劳动女性人口仅略多于印度，但其女性劳动力总规模却是印度的近三倍。

现金具有无可辩驳的说服力。当女性能够挣得可观的报酬时，家里的男人们就不得不重新审视“女人干不成大事”的成见。“过去，连我亲哥哥都不信任我，”利雅得一家美容院的老板瓦法·奥泰比说道。她曾向哥哥借钱开办家庭彩妆工作室，哥哥借了钱，却嘲笑她根本不知道自己在做什么，还说做这个很丢人。然而当她在六个月内悉数还清借款后，哥哥甚至主动提出想入股她蒸蒸日上的生意。“男人们对待女性的态度变了，”奥泰比表示，“他们现在对女人的能力感到不可思议。”

工资同样激励着女性自身，尤其是当她们能够自主支配这些收入时。杜克大学埃里卡·菲尔德等人的一项研究表明，在一项公共工程项目中，如果将薪酬直接存入印度女性自己的账户而非其丈夫账户，并对她们进行账户使用培训，她们的工作积极性会显著提高。

金钱也在改变富裕阶层的传统偏见。帕尼塔·阿加瓦尔是北方邦一家小型工厂老板的女儿，她回忆起自己的成长过程中，充斥着丈夫专横跋扈、妻子被要求在晚上七八点前必须回家的家庭氛围。她暗下决心一定要自立赚钱，以便“万一陷入受虐待的婚姻，我买得起自己的房子，能毫不犹豫地抽身离开”。

当她获得英国曼彻斯特大学的录取通知时，部分亲戚劝阻她的父母不要放行。他们认为这笔钱不如留作嫁妆，毕竟女孩的归宿就是成家；他们还警告称，出国留学将使她变得难以管束。

但她依然毅然成行。当她学成归国，满脑子都是如何推动家族生意发展、让父母财富增值的点子时，原先持反对态度的亲属终于意识到他们错了。阿加瓦尔说，现在那些亲戚也纷纷将自己的女儿送去深造了。

图片来源：盖蒂图片社 (Getty Images)

互联网连接正在加速这种变革。阿富汗科技界女性先锋罗亚·马赫布布写道，在极度男权至上的社会中，数字技术让女性能够“憧憬另一种完全不同的人生”。往往最能打动人心的范例并非来自遥远而陌生的极度自由化国家，而是来自那些思想略显开放的近邻。沙特阿拉伯的人力资源经理纳瓦尔·奥纳齐说，她年轻时受到来自其他阿拉伯国家的卫星电视节目的深刻启发。她在屏幕上看到了担任要职的职场女性专访，看到了描绘相对自由平等家庭生活的埃及喜剧。最终，她与那个禁止她外出工作的丈夫离了婚。

科技也能潜移默化地动摇顽固家庭。达特茅斯学院的玛德琳·麦克尔韦在北方邦的一项研究表明，仅需向丈夫和公婆播放一段时长六分钟的视频——内容是女性分享参加职业培训项目的真实体验——就能使妻子获得工作的概率大幅提升35%至57%。该视频特别强调了工作场所（一家纺织工坊）安全且全为女性员工，而工作的女性依然有充裕时间操持家务。

科技还助力女性精准获取就业机会。通过网络招聘平台，女性足不出户便可浏览岗位并投递简历，部分岗位甚至支持远程居家办公。浙江大学陆菁的一项研究发现，数字技术的普及通过降低求职门槛并创造出诸如在线客服等契合女性的全新岗位，显著提升了中国女性的就业概率。

此外，社会从众效应也在发挥作用。人是社会性动物，绝大多数人在面对道德伦理问题时并非从第一性原理推导答案，而是观察同伴的行事法则。芝加哥大学莱昂纳多·伯斯汀在2020年的一项研究中发现，大多数沙特已婚年轻男性实际上赞成女性外出工作。然而，四分之三的人都严重低估了持相同看法的同胞数量。当被告知真实的统计数据后，他们更倾向于主动协助妻子找工作。

利雅得美容沙龙老板奥泰比指出，即便是保守传统的沙特男性，如今也“开始习惯身边的女同事……这彻底重塑了人们的择偶观。这太令人惊喜了！人们现在是通过职场结识后恋爱结婚，而不是依赖包办的传统方式。”

乔治梅森大学的约翰娜·莫勒斯特罗姆指出，女性从事的具体职业类型至关重要。如果已婚女性所能接触到的工作大多既危险又脏乱，那么妻子外出打工在邻里或教会眼中，便等同于在公开宣扬其年富力强的丈夫是个无能的懒汉。随着国家日渐富足、工作环境更加体面舒适，这种污名化标签自会烟消云散。

埃文斯博士换了一个角度解释：如果女性放弃工作的机会成本上升（得益于就业前景改善），或者女性外出工作的道德压力与舆论成本（所谓的“蒙羞”）下降，那么“荣誉与收入的权衡取舍”就会被彻底颠覆。沙特领导层一直在协同调动这两大杠杆：一方面强制推行雇佣沙特本国国民的企业配额制，另一方面高调释放“女性抛头露面参与社会是体面之事”的社会信号，例如王室女性公开参与时尚大秀等。

挣脱沉重枷锁

摆脱父权禁锢最彻底的方式，或许莫过于迁移到一个不再受旧习俗束缚的地方。在萨拉·穆罕默德17岁时，她的哥哥曾用一把卡拉什尼科夫步枪顶着她的头，逼迫她接受一场“换亲”。他逼她嫁给一个从未谋面、年龄比她大一倍的男人，以换取她哥哥能够迎娶该男子的妹妹。当时身在伊拉克的穆罕默德接连拒绝了两次。她哥哥威胁称，如果她说第三个“不”，他杀她“就像喝口水一样容易”。她只能含泪答应。

然而，就在婚礼紧锣密鼓筹备之际，她借来一件罩袍乔装潜逃，躲到了同情她的表亲家中，最终辗转在瑞典获得了政治庇护。如今，她主持着反对基于荣誉的暴力慈善机构GAPF——并拥有了一位无论做任何事都无需事先“报批”的伴侣。

然而，移民放弃陈旧传统的过程绝非一帆风顺。女性极少能够独自迁徙，往往是与男性亲属同行，而这些男人往往千方百计将她们与当地主流社会隔离开来。此外，带有极端“荣誉文化”背景的族群涌入自由社会，往往引发本土居民的抵触。例如，许多瑞典人认为，十年前来自穆斯林国家的庇护寻求者激增，将侵蚀本地的女权主义价值观。这已成为欧美民粹主义右翼极具煽动力的政治号召。

然而，这类恐慌言过其实。本土居民极少会被新来者的保守观念同化，反而是移民群体展现出强烈的融入趋势。在美国，来自女性就业率极低国家的移民群体，其第二代与本土女性的就业率差距已基本抹平。一项针对从46个国家迁往18个欧洲国家的移民研究表明，第二代女性移民比男性移民持有更为平等的性别观念；且其祖籍国的男女越不平等，第二代女性思想观念的跨越幅度反而越大。

自由文化具有天然的吸引力，对女性尤甚。幼年跟随家人从土耳其移居瑞典的基内姆（化名）暗下决心，绝不过那些表姐妹们“复制粘贴”般的人生——她们“不读书深造”，“早早嫁人”，而且“住着格局一模一样、陈设毫无二致的公寓”。

她的反叛是循序渐进的。她的父母并非无可救药的顽固派，他们允许她读大学，但要求她每个周末必须回家。她逐渐减少了回家的频次，并开始参加派对，尽管她从不饮酒。凭借化学工程学位，她在医疗器械行业找到了一连串不错的工作，并在斯哥尔摩购置了自己的公寓。这让她的父母大为不悦，但她毫不退让。最终，她嫁给了一位瑞典人。这彻底击穿了她父母的承受极限，并与她断绝关系长达五年。

后来他们冰释前嫌，如今父母已全然接受她能够自主选择人生。今天，基内姆说她大约有一半的表姐妹都在工作。“她们终于意识到，自己也能为家庭赚来真金白银。而没有人会跟钱过不去。”■

ch029.xhtml

乌克兰战争愈演愈烈、战线外溢且深陷泥潭

作为乌克兰盟友的瑞典悲观研判：已做好与俄罗斯对抗50年的心理准备

插画摄影：克洛伊·库什曼 (Chloe Cushman)

2026年9月1日

疲惫不堪的国际社会或许并不愿面对这一事实，但乌克兰战争正在不断加剧，而非步向终局。随着冲突走向深水区，战火与外溢风险也在持续扩大。北欧各国秉持严肃务实态度的政府，正越来越多地未雨绸缪，着手制定应对未来数十年来自俄罗斯威胁的长期国防规划。

繁荣而审慎的瑞典便是典型范例。仅仅在两年前，在普京军事侵略的驱使下，这个北欧国家彻底放弃了长达两个世纪的中立传统（该传统一度演化为近乎天真的绝对和平主义），毅然加入北约。随着瑞典选民将于9月13日走向投票站，支持乌克兰已成为全国高度一致的共识，以至于选战辩论中几乎鲜少提及战争议题。所有主要政党——包括曾对俄罗斯暗送秋波的民粹主义左右翼阵营——全盘支持对乌克兰提供经济与军事援助，对弗拉基米尔·普京总统抱有根深蒂固的不信任，并一致认同必须建设坚不可摧的国防体系。

在斯德哥尔摩，令人惊讶的是政界高层频繁使用“幼稚”一词，来反思国家迟至2000年代末对俄罗斯所持的绥靖与幻想策略。左右两派的风云人物皆竭力表明，他们如今有多么冷酷而务实。

在目前的保守派联合政府执政下，瑞典直言不讳地指出俄罗斯实为外强中干，指责普京及其心腹利用虚假的统计数据掩盖其千疮百孔的财政状况，并借残酷的政治高压压制国内的异见声浪。然而，瑞典首相乌尔夫·克里斯特松并不认为普京的狂热好战情绪短期内会退潮。相反，在8月27日接受《经济学人》旗下视频节目《地缘政治内参》（Inside Geopolitics）专访时，克里斯特松表示，他“根本看不到任何迹象表明俄罗斯改变了发动这场战争的底层逻辑。从根本上说，他们就是要征服乌克兰，而我们根本无法预料他们达成之后还会有什么野心和目标。”

克里斯特松承认，乌克兰同样在承受着与日俱增的剧痛：由于防空拦截导弹短缺日益严重，各大城市暴露在俄罗斯巡航导弹与喷气动力无人机的狂轰滥炸之下——他特别指出，中东战火消耗了大量原本可用的防空武器库存，使这一危机雪上加霜。但他随即正告俄罗斯，不要妄想乌克兰的抵抗意志会瓦解，坚称该国展现出了惊人的战略韧性。

在斯德哥尔摩看来，两股不可阻挡力量的正面碰撞——俄罗斯的新帝国主义野心与乌克兰的爱国主义情怀——彻底铸成了一场僵局。战事的代价正在节节攀升，但对交战双方而言，战败屈服是更难承受的梦魇。这位首相坦言，自己无法预判战争将于何时落幕。“但战争以何种方式收场，将至少决定我们所在地区未来一整代人的安全格局，”他说道。

为做好周全准备，他的政府全力将瑞典打造成乌克兰的长期防务与工业伙伴，向乌克兰空军捐赠并出售“鹰狮”（Gripen）战斗机。瑞典正在扩大义务兵役制规模，并深化与北欧、波罗的海邻国，以及英、法、德、荷、波等国的防务安全联盟。

在中左翼阵营，对俄防务立场同样毫不妥协，而民调显示该阵营极有可能在当月的选举中获胜。在冷战时期曾高调反对加入北约的社会民主党，如今其核心领导层公开表示，政界已形成跨党派共识，即认定俄罗斯在可预见的未来是“长期的生存威胁”，瑞典必须拿出持续50年的决心，全方位强化军事与社会维度的国家全面防御。

在斯德哥尔摩，人们对乌克兰战争的结束既翘首以盼，又心存忧虑。高层政要普遍担心，迄今表现令人惊叹的欧洲团结将面临严峻考验，因为那些距离俄罗斯较远的国家势必会主张与普京实现关系回暖与正常化。社会民主党外交事务发言人、未来可能出任外交大臣的摩根·约翰松担忧，停火协议带来的不过是乌克兰冲突的“彻底冻结”，绝非长治久安的和平。当被问及瑞典是否还能信任俄罗斯时，他设想了一个极其遥远的未来：俄罗斯通过政权更迭迎来戈尔巴乔夫式的改革派领导人。约翰松指出，即便是到了那时，瑞典人“也必须枕戈待旦保持警惕”，切不可重蹈25年前的覆辙——当年政客们不再警惕俄罗斯的扩张野心，从而亲手拆解了本国的防御能力。

瑞典是乌克兰在取得反腐与治理成效后加入欧盟最坚定的呼吁者之一，主张借此将乌克兰牢牢锚定在西方阵营并捍卫其民主制度。然而，包括法国在内的其他大国虽然支持乌克兰，却对欧盟盲目东扩充满戒心；一些农业人口比重较大的欧盟成员国，则深切担忧乌克兰粮食农业涌入带来的冲击。

既无乐观余地，亦无替代选项

上述利益分歧使瑞典成为观察围绕俄乌战争的西方团结脆弱性的绝佳窗口，尤其是在唐纳德·特朗普可能试图逼迫乌克兰接受一份不公正、缺乏安全保障的屈辱和平之际。瑞典一直试图与特朗普保持良好关系。但私底下，官员们深感焦虑：如果一心不惜代价求和的美国禁止盟友向乌克兰转运美制武器及零配件、掐断关键的战场情报支持，甚至威胁从波罗的海国家或波兰撤军，局势将会何去何从？斯德哥尔摩的政客们寄希望于美国深知这将付出不可承受的政治代价。一位资深人士甚至暗示，特朗普眼下或许已经对乌克兰失去了兴趣，转而聚焦于伊朗与中国。

乌克兰的盟友们很难看清战争体面收场的明确路径。他们既害怕一场无论对乌克兰还是对欧洲安全而言都极度不利的“恶性妥协”，又对俄罗斯抱有彻底且持久的戒心。因此，唯有僵局持续：在这个残酷的现实面前，一个糟糕的结局，竟好过了所有其他更坏的可能。■

ch030.xhtml

黄仁勋何以成为人工智能界的“泰勒·斯威夫特”？

当这位英伟达掌门人回到故土台湾时，周围的一切便陷入某种近乎狂热的躁动

2026年9月3日

每年六月的某几天，台北便会陷入一种特殊的亢奋状态。得知黄仁勋可能在某家餐厅用餐的消息后，数百名民众顶着酷暑聚集在餐厅门外。每隔几分钟，一辆疾驰而过的黑色轿车就会引发一阵骚动——大家猜测，他或许就坐在车里。

作为全球市值最高公司的掌门人，黄仁勋每年夏天都会回到出生地台湾，参加亚洲规模最大的电脑硬件盛会——台北国际电脑展（Computex）。在台期间，他通常会溜到街头小吃摊吃夜市，四处闲逛并与路人寒暄。今年的一天晚上，他亲切地穿行在饶河街夜市，品尝抹茶甜品，还跟烤玉米摊的老板开玩笑说，只要允许他插队，他就请排队的所有人吃晚饭。大批台湾民众簇拥在他身后，希望能与他自拍、索要签名，或是闲聊两句。当晚晚些时候，黄仁勋在一家餐厅卫生间的瓷砖上写下了“仁勋到此一游”（Jensen was here）。一瞬间门外便排起了长龙：人人都争相为这件珍贵的“仁勋墨宝”拍照留念。

在另一个晚上，黄仁勋与几位韩国科技公司的掌门人一同前往餐厅。当他跨出车门的那一刻，人群高举手机蜂拥而上，安保人员拼尽全力才勉强阻挡住汹涌的人潮。随行的科技高管们一脸错愕，仿佛误入了一场不属于自己的狂欢活动。相比之下，黄仁勋却显得泰然自若。他微笑着挥手，从容地配合镜头拍照，随后步入室内。“他的皮肤怎么这么好，”站在我身旁的一位粉丝叹道。人群中甚至有人身穿印有黄仁勋头像的T恤、挂着他的脸谱钥匙扣——这些周边在会展中心卖得火热。

“他的皮肤怎么这么好，”站在我身旁的一位粉丝叹道

在人工智能淘金热中，黄仁勋就是那个“卖铁锹的人”。他联合创立的英伟达设计并销售着性能超群的芯片，为全球约三分之二的AI算力提供动力。全球最大科技巨头的掌舵人以及争相建立本国自主AI算力基础设施的各国政府，无不对他礼遇有加，渴望获得英伟达供不应求的处理器。资本市场对他更是顶礼膜拜：这位身着标志性皮衣的英伟达老板一句顺带的赞誉，就能让一家公司的股价一飞冲天。

然而，仅凭这一点并不能完全解释他在台北所引发的狂热。马克·扎克伯格和萨姆·奥尔特曼固然是权倾一时的科技领袖，但很难想象美国人会排长队只为一睹他们的真容，或是购买印有他们头像的文化衫（扎克伯格曾评价黄仁勋“就像科技界的泰勒·斯威夫特”）。

在台湾，这种被称为“仁来疯”（Jensanity）的狂潮，部分源于对本土英雄的崇拜。出生于台湾、幼年移民美国的黄仁勋，是台湾历史上最成功的商业传奇。他还为台湾的电子产业做出了无可比拟的全球公关，盛赞台湾是这场AI革命的“中心枢纽”。全球绝大多数尖端芯片都由台湾积体电路制造公司（台积电，TSMC）代工生产。在这个仅有少数国家在官方层面予以承认的岛屿上，政界人士希望台湾在AI浪潮中的核心枢纽地位能够转化为强大的软实力。这份软实力至关重要：随着来自中国大陆的姿态日益强势，台湾的安全在很大程度上寄托于外部力量的协防意愿——尤其是美国。

黄仁勋不仅提升了台湾的知名度，还让许多台湾人身家暴涨。“仁勋极大地拉升了我的投资收益，”一位在餐厅外守候黄仁勋的年轻工程师告诉我。台湾基准股票指数在过去一年内上涨了90%，一跃成为全球第五大股票市场。这一增长在很大程度上归功于为英伟达代工芯片的台积电的惊人涨幅。自2023年以来，这家台湾企业的市值翻了五倍多，达到约2万亿美元，其旗下的数十家本土供应商也随之大获其利。

然而，正如我在近距离跟拍黄仁勋的三天里所观察到的，这种狂热很大程度上也是由这位英伟达掌门人亲自推波助澜的。与其他硅谷科技巨头不同，黄仁勋极力拉近与普通公众的距离。一天晚上在台北，当他与父母在一家餐厅用餐时，他特地端出几盘热气腾腾的包子，分发给守候在外的媒体记者。

这种亲民的人设对其商业运营裨益良多。分析师指出，黄仁勋宛如摇滚巨星的地位，让台湾制造商争相寻求合作，巴望着能分得他的一缕星光。在AI供应链极其紧俏、芯片设计公司四处求托代工厂而非代工厂争抢订单的当下，这为英伟达带来了无可比拟的竞争优势。在今年的电脑展结束后，黄仁勋旋即飞往首尔，在啤酒与炸鸡的酒桌上拉拢韩国存储芯片企业的掌门人，甚至被当地电视台拍到随着韩国流行音乐稍显笨拙地起舞。台湾民众送给黄仁勋“全民干爹”的亲切绰号，绝非空穴来风。

近二十年前，当我以一名芯片设计工程师的身份首次参加Computex等硬件展会时，那里的氛围可谓枯燥乏味。不善社交的工程师们扎堆检视服务器机架、研讨散热指标、敲定零部件价格。当时媒体关注的焦点是微软和谷歌，而为这些科技巨头设备提供底层动力的台湾芯片制造商，在业内圈子之外几乎籍籍无名。

迄今为止，黄仁勋维持了错综复杂供应链的顺畅运转。但其他严峻挑战正在步步紧逼

然而，训练前沿AI模型需要极其庞大的计算基础设施，过去几年里，硬件制造商被猛然推到了聚光灯下。本届电脑展的火爆气息扑面而来，绝不仅限于穿梭在展馆上空的无人机。昔日记忆中零落沉闷的展位，被巨型霓虹灯展馆所取代；人形机器人在通道间四处穿梭，参展商们热情地分发帆布包和珍珠奶茶。电脑展如今迎来了众星捧月的主角：为了挤进大会议厅聆听黄仁勋的主题演讲，人们提前数小时便排起了长队，队伍延绵数千人之多。

当他终于身着标志性的黑皮衣登台亮相时，全场掌声雷动。黄仁勋以向台湾的供应商与合作伙伴致谢拉开帷幕，他身后屏幕上闪烁着各大知名电子制造商的徽标，预示着他们即将被逐一致意。然而，他第一个公开致谢的“供应商”，居然是通化夜市的一位水果摊贩；紧随其后的是一家本土生煎包店，然后是一长串餐饮店家。台下的观众对此报以热烈的欢呼与掌声。

黄仁勋走向全球经济舞台中央的历程并非循规蹈矩。他于1963年出生于台湾，九岁时与哥哥一同被送往美国投奔叔父。叔父误将肯塔基州的一所不良少年感化寄宿学校当成了贵族菁英名校，阴差阳错地将兄弟俩送了进去。黄仁勋睡在集体宿舍，还被安排去打扫公厕，但因年龄太小无法在校内上文化课。于是他前往当地公立学校就读，那里的孩子因他的口音对他百般欺凌，甚至蔑称他为“中国佬”（chink）。黄仁勋将那段艰难岁月视为磨砺自己坚韧意志的宝贵财富。

兄弟俩赴美两年后，父母飞抵美国与他们团聚，全家在俄勒冈州安家落户。黄仁勋随后攻读了电气工程专业，随后南下闯荡硅谷。1993年，他与另外两位工程师联手创立了英伟达。日后让英伟达声名大噪的核心产品图形处理器（GPU），在当时主要用于电子游戏领域，且赛道竞争极度白热化。英伟达凭借研发出性能更优越的芯片，更重要的是抢先一步推向市场，从而在竞争中脱颖而出。

到了2010年代中期，黄仁勋坚信AI将重塑整个计算体系，而擅长同时处理海量并行计算任务的GPU，注定会成为AI的核心基石。在很长一段时间里，这都是一种离经叛道的冷门观点。AI虽然一直被鼓吹为“下一个风口”，但那个颠覆性的时刻似乎始终遥遥无期。直到2022年11月，模型开发商OpenAI发布了ChatGPT。用于训练ChatGPT的英伟达芯片需求随即迎来爆炸式激增。短短三年内，英伟达的市值从3400亿美元一路狂飙至惊人的5万亿美元。

斯蒂芬·班农曾公开斥责黄仁勋是“影响美国利益的代理人”

公司的成功不仅依靠绝妙的芯片架构设计，同样取决于驾驭供应链合作伙伴关系以及应对牵扯各方的复杂政治环境的能力。软件公司的领袖往往被塑造成孤军奋战的独行天才，但硬件制造在本质上是一项庞大而精密的协同工程。英伟达的最新旗舰产品Vera Rubin NVL72，是一套专为训练和运行超大规模AI模型而打造的完整超算系统，内部包含130万个零部件、1000多颗芯片以及与之配套的复杂控制软件。打造这样一套设备，需要数以百计的供应商、代工厂和软件公司紧密协作。

AI算力狂潮引发了对这些零部件的激烈争抢，而黄仁勋的核心工作之一，就是确保各大供应商将英伟达的订单列为最高优先级。雄厚的财力在此发挥了关键作用：英伟达利用其强大的资产负债表与制造厂商签署了多年长期采购协议。不仅如此，黄仁勋还需要供应商加速扩大产能。分析人士透露，黄仁勋曾多次向台积电施压，力促其摒弃在投资新建产能上的保守态度。

迄今为止，黄仁勋维持了错综复杂供应链的高速运转。但其他挑战正接踵而至，尤其是来自英伟达最大客户的竞争威胁。Alphabet、亚马逊、Meta和微软都在竞相研发自研AI芯片，以摆脱对英伟达的过度依赖。与此同时，英伟达的股价已被视作衡量资本市场对AI热潮可持续性信心的风向标，并在今年内多次出现回调震荡。黄仁勋的应对之策是将英伟达与AI的未来更加牢固地绑定在一起。他不再仅仅满足于兜售底层芯片，而是要统揽驱动整个行业运转的底层基础设施、软件生态系统和AI模型。

这种战略转型需要更加深厚的供应链公关与外交手腕，因此看到黄仁勋在展会期间穿梭于各展厅与制造合作伙伴打成一片，我并不感到意外。他的随行人员透露，黄仁勋之所以坚持探访供应商展位，是因为他的现身能聚拢庞大人流，为伙伴们带来极大的曝光度。伴随着穿透展厅的“Jensen！Jensen！”欢呼声，他穿梭于各展台之间，与一线工程师和企业高管畅聊，不时与围拢的人群击掌互动。直到傍晚时分，他才露出一丝倦容，在一家展位的地板上盘腿而坐，与一位台湾高管喝起了啤酒。数百名狂热粉丝瞬间将其团团围住，纷纷举起手机记录这一瞬间。

发表完主题演讲后，我与黄仁勋在展厅的一间贵宾室相对而坐。彼时他已脱下那件标志性的皮夹克，换成了一件藏青色纯棉外套。门外，随行团队与安保人员正行色匆匆地忙碌穿梭，而室内的黄仁勋则显得波澜不惊、气定神闲。

当我解释说自己更想探寻他个人的内心世界而非仅仅聚焦英伟达时，他显得有些羞怯。他坦言自己并不习惯谈论个人私事，并提前致歉表示如果回答得不够出彩还请见谅。在舞台上，他是一位极富感染力与激情的演讲大师；但在私下交谈中，他的语调几乎低沉如耳语，语速沉缓且字斟句句。

我询问自ChatGPT横空出世以来，他的个人生活发生了怎样的改变。硅谷的发展史充斥着一夜成名的创始人神话。而黄仁勋的崛起足足经历了三十年的积淀与蛰伏，这或许使得当巨星光环最终降临时，那种人生巨变显得更加光怪陆离。他坚称，自己的一切并未发生本质改变。他一如既往地穿衣打扮，正如他自己所言，每天早上起来依然“一只脚一只脚地穿裤子”。

分析师指出，黄仁勋宛如摇滚巨星的地位，让台湾制造商争相寻求合作

黄仁勋不仅要适应自己全球闻名的公众身份，还发现自己置身于中美地缘政治博弈的风口浪尖。自2022年10月以来，美国政府一直在竭力遏制其竞争对手获取先进计算芯片。与此同时，英伟达在台湾的各大供应商也笼罩在潜在军事对抗升级的阴影之下。黄仁勋突然发现，自己必须学会与各国首脑建立通畅的互信纽带，尤其是唐纳德·特朗普——黄仁勋一直在游说特朗普放宽对华芯片出口限制。

他表示，与政界政客周旋并非其天性使然。尽管如此，他依然展现出了非凡的政治手腕，善于将商业诉求转化为打动政客的语言。他将庞大的数据中心重新包装为“AI工厂”——这一精准措辞极大契合了特朗普提振本土制造业的偏好。而针对那些对依赖美国跨国公司心存戒备的外国政府，他顺势提出了“主权AI”的概念——即依托本国价值观构建、在底层由英伟达强力驱动的国家级自主基础设施。

今年5月，当他随同特朗普出访中国参加与国家主席习近平的峰会时，其外交手腕展现得淋漓尽致。黄仁勋似乎是在最后一刻被临时增补入访问团的——当特朗普在“真实社交”（Truth Social）上宣布“伟大的黄仁勋”正随同搭乘“空军一号”时，外界无不深感意外。在访问北京期间，黄仁勋与中国高层官员举行会谈，并欣然接受了中国顶尖学府清华大学顾问委员会的席位任命。

对华鹰派人物对此深感疑虑。前白宫首席战略师斯蒂芬·班农斥责黄仁勋是中国共产党的“影响力代理人”，并公开呼吁将其逮捕。来自马萨诸塞州的民主党参议员伊丽莎白·沃伦指出，黄仁勋游说向中国出售尖端AI芯片的行为，可能会“为中国的军事发展注入涡轮增压，并削弱美国在科技领域的全球领导地位”。

然而，令人颇感微妙的是，对于一个被扣上“对华绥靖”帽子的商人而言，黄仁勋同时也是台湾坚定的公共声援者。他直言是台湾成就了英伟达，并称台湾是英伟达的“商业故乡”。英伟达正在台北兴建其在美国之外规模最大的办公大楼之一，并计划每年在此投资约1500亿美元。这种钢丝上的平衡并非总是顺风顺水——2024年，黄仁勋因在公开场合称台湾为“世界上最重要的国家之一”而引发轩然大波（随后他被迫紧急澄清，强调自己旨在赞扬台湾在全球电子供应链中的重要地位，而非在政治上承认其为主权国家）。在电脑展上，他苦笑着对我感慨，在当今的时代背景下担任AI硬件帝国的领军人，无异于在承担“有史以来最具挑战性的沟通工作”。

黄仁勋调动供应商遵从其意志，绝非单凭个人魅力。曾与黄仁勋共事数十年的台积电前高管蔡力行透露，黄仁勋在工作中“极具进取心”且“雷厉风行”。蔡力行将黄仁勋处理供应链伙伴关系的方式形容为一种“粗犷的正义”：双方心照不宣地明白，有时某一方在理，有时某一方出错，但从长远合作来看，彼此的得失大致能维持平衡。

黄仁勋不仅提升了台湾的知名度，还让许多台湾人身家暴涨

这种“粗犷的正义”同样体现在英伟达内部的治理风格中。内部员工透露，黄仁勋对推诿借口缺乏容忍度，一旦项目进展受阻，他会当场毫不留情地提出质疑。有大约60位高管直接向他汇报，这在跨国企业首席执行官中极为罕见。他通常会当众直截了当地给出反馈意见，其逻辑在于让会议室里的所有人都能从挫败中吸取教训。

我有幸偶尔目睹过这种不加掩饰的风格流露。在一场新闻发布会上，一名员工试图迅速夺下一位分析师手中的麦克风，黄仁勋当即对其予以训斥。当一名BBC记者追问台湾面临的地缘政治风险时，黄仁勋起初给出了滴水不漏的标准回应；而当该记者继续穷追不舍时，他半开玩笑地调侃说，现场应该有人帮BBC想几个新问题。全场哄堂大笑，但黄仁勋脸上的不悦之色清晰可见。

在英伟达任职逾二十年、现任副总裁的资深工程师列夫·勒巴雷迪安表示，在局外人看来，黄仁勋的高管例会氛围可能“令人窒息”。会上通篇探讨的都是遭遇的棘手难题与失败教训，鲜有报喜的好消息。

但勒巴雷迪安指出，黄仁勋在公司内部构建了一种文化，员工深知错误虽会被当面指出，但绝不会因此遭受恶意惩戒。英伟达的核心高管团队在过去二十年间保持了惊人的稳定性，这在动辄剧烈动荡的科技行业极为罕见。高管们同样获得了丰厚的回报。黄仁勋常常自豪地宣称，自己培养出的亿万富翁高管人数超过“全球任何一位CEO”。基层普通员工的待遇也同样水涨船高。

当我问及他的内驱动力究竟是什么时，黄仁勋并未高谈阔论殖民火星、克服死亡或重塑人类未来。他淡淡一笑，指出埃隆·马斯克是他的好友，但两人的追求截然不同。他说，自己的目标要纯粹得多：他只希望自己的员工和合作伙伴都能过得好。“也许，”他微笑着说道，“我只是不够雄心勃勃吧。”■

沙勒什·奇特尼斯 (Shailesh Chitnis) 是《经济学人》全球商业报道记者。

摄影：刘仲良 (Leong Chong Lao)

ch031.xhtml

会有人像程序员那样高频使用AI吗？

这一问题的答案，将对整个AI投资繁荣产生深远影响

插画：西蒙·贝利 (Simon Bailly)

2026年8月30日

过去几年里，硅谷初创企业中软件工程师的日常工作模式发生了翻天覆地的变化。最终的产出依然未变：依旧是海量的代码行。但如今他们不再整天坐在屏幕前敲击键盘，而是对着麦克风低声耳语。他们的语音指令随后通过人工智能被转录，并输入给AI智能体（AI Agents）——这些能够与软件交互、并在此场景下直接编写代码的数字助手。

ChatGPT母公司OpenAI的科林·贾维斯坦言，其他白领群体目前尚未“踏入这一全新世界”。而他所在公司的未来——乃至整个人工智能狂潮的走向——都维系于这种局面能否被彻底打破。为了证明在数据中心领域天文数字般的巨额资本开支合理，美国的云计算巨头以及Anthropic和OpenAI等前沿AI实验室，必须确保该技术的普及步伐继续飞速狂飙。《经济学人》的初步粗算显示，为了完全覆盖数据中心建设与运营的庞大成本，到本年代末，全球由AI衍生带来的年总收入必须从目前的约1500亿美元激增至每年2.5万亿美元。

迄今为止，AI应用渗透率最高的领域毫无疑问是软件开发。根据知名开发者在线技术社区Stack Overflow的调查，五分之四的程序员表示自己在日常开发中使用AI编程工具。调研机构SemiAnalysis测算，对于Anthropic和OpenAI而言，编程场景贡献的收入如今已占到其年度经常性收入（ARR，一种根据最近单月收入年化推算的流行营收指标）总和的一半以上。2025年6月，Cognition、Cursor、Lovable和Replit这四家AI编程初创企业的ARR总和约为8亿美元；而今，这一数字已暴增至60亿美元。

图表来源：《经济学人》

程序员群体的空前热情，完美解释了为何在软件开发人员占比越高的行业，AI技术的渗透与扩散速度就越迅猛（见图表）。乐观派可能会将此视为有力佐证，证明AI浪潮才刚刚拉开序幕，随着其他领域应用的不断成熟，未来数年内围绕该技术的资金投入必将攀上全新高峰。然而，编程能够成为一枝独秀的特例，背后可能存在不可忽视的深层结构性动因。

常被寄予厚望成为AI下一批颠覆性爆发点的三个领域分别是：法律服务、金融业和客户服务。这三大领域的增长势头同样令人瞩目。Harvey、Clio和Legora这三家AI法律工具开发商，其ARR总和在过去一年内翻了一番，攀升至10亿美元。Harvey透露，律师在其平台上停留的使用时长正以每月翻倍的速度狂飙。银行家及其他金融专业人士也开始大举拥抱这项技术。为金融服务业打造AI专业工具的初创公司Rogo，在上一季度新增了100家大型企业客户，ARR激增了50%。

AI在客户服务领域的应用同样呈燎原之势。专门深耕该赛道的Sierra公司在6月份实现了2亿美元的年化经常性收入，较去年11月翻了一番。风险投资界对此极度看好：今年迄今，风险投资机构已向该领域的AI初创公司注资30亿美元，远超其他任何AI垂直应用门类。（相比之下，对AI编程初创公司的投资节奏有所放缓，部分原因在于前沿AI实验室自研的原生编程工具迭代扩张过于迅猛。）

这三类应用场景与编程有着诸多相似之处：它们都拥有广阔的潜在市场空间，且业务流程往往具备极强的规范性与套路化特征。然而，有四大独特维度将编程与其他行业截然区分开来：海量训练数据的易得性、模型输出验证的低门槛、工作流程本身对人际交往互动的极低依赖，以及软件工程师独特的群体画像。AI企业正试图在其他行业弥合这些鸿沟，但这绝非易事。

AI之所以能成为无可替代的优秀编程搭档，部分原因在于互联网上公开沉淀了浩瀚如烟海的开源代码库。这些代码因此顺理成章地被喂入模型，在某些特定模型的训练集中，代码数据甚至占到了全部训练总量的近五分之一。然而，针对繁琐的客户投诉处理或复杂的商业合同谈判，公网空间根本不存在体量如此庞大且开源共享的文本宝库。

此外，AI生成的代码相对而言极易通过自动化手段进行核验验证。软件测试本就是现代软件工程的标准规范流程：一套代码逻辑可以直接自动化检验另一套代码的准确性。这大幅提升了在AI代码引发系统崩溃前及时捕获缺陷的概率。不仅如此，测试反馈通过在模型出错时发出明确的报警信号，有助于持续自我强化并反哺模型性能的跃升。相比之下，评估一位怒气冲冲的客户是否得到了真正抚慰，或是某套估值模型是否精准捕捉了企业财务报表中的微妙语境，难度要高出几个量级。

开发代码生成工具还拥有得天独厚的优势：完成一项开发任务所需的大多数关键信息，早已规范地沉淀于软件的底层代码库之中，或者可以通过连接不同软件的数字化管道——应用程序编程接口（API）轻松获取。然而在许多其他现实业务场景中，最致命的核心信息根本无法以结构化数字形式留存，它们要么深埋在呼叫中心话务员的脑海深处，要么隐匿在谈判对手律师难以捉摸的面部表情之中。

AI公司正在全力攻坚这些行业硬伤。法律AI领军企业Harvey的掌门人温斯顿·温伯格透露，伴随着技术跃进，其公司如今已能部分依赖由AI系统自主生成的“合成数据”来定向训练其行业模型。诸如Mercor之类的初创公司，正在将AI企业与各行各业的顶尖专业人士对接，由后者将独家隐性行业知识转化为模型可学习的资产。AI客服初创公司Cresta创始人吴平表示，通过深度挖掘并剖析数以万计的呼叫中心沟通录音与文本记录，其系统能够推断并沉淀员工脑海中的大量业务诀窍，例如复杂的退款异常流程该如何权衡处置。大批所谓的“前线驻场工程师”（forward-deployed engineers）正被AI企业派驻到企业客户的现场办公室，实地排查痛点并现场迭代定制工具。

然而，AI在编程领域大放异彩的最后一个核心支柱，恐怕最难在其他领域复制：那就是软件工程师群体本身。作为一群既带有极客特质又对技术变迁习以为常的人（毕竟编程语言和技术栈每隔几年就会迭代更替），他们是新技术天然的敏锐拥趸与尝鲜者。顶级风投机构安德森·霍洛维茨（Andreessen Horowitz）的阿尼什·阿查里亚指出，AI在编程领域的普及浪潮呈现出鲜明的“自下而上”路径，工程师们完全是自发地挖掘、尝试并引入这些前沿工具。

反观呼叫中心的话务员，他们通常既无动力、也无权限在日常流水线作业中擅自尝试新工具，往往只能被动等待自上而下的系统部署命令。而律师和金融从业人员，则时刻处于严密细致的职业合规与法律红线监管之下，任何业务创新都必须如履薄冰、谨小慎微，以免触犯行业监管雷区。

其他垂直领域的AI应用固然会继续稳步演进，且伴随着底层模型架构的持续迭代以及企业对向AI智能体放权的顾虑逐渐消解，落地速度或许还会显著加快。但要想让绝大多数办公室白领也像程序员那样，整天对着麦克风低声耳语来发号施令，恐怕尚需漫长时日。■

ch032.xhtml

像CoreWeave这样的新型云厂商规模剧增，风险亦然

支撑其急剧扩张的债务高山正在迅猛堆积

图片来源：盖蒂图片社 (Getty Images)

2026年9月3日

全新颠覆性技术的横空出世，往往会为各类中间商提供丰沃的滋养土壤。1960年代，精明的商人从IBM大批量扫货大型机，转手租赁给银行和军工国防企业。在互联网泡沫时代，“域名倒爷”低价抢注网站域名并高价倒手套利。人工智能时代同样概莫能外。以CoreWeave为代表的一批“新型云厂商”（neoclouds），通过大举囤积采购专用AI算力芯片并向外出租算力，一跃成为整个AI产业链中扩张最为迅猛的黑马。然而，它们同时也是抗风险能力最为脆弱的一环。

新型云厂商在过去几年的强势崛起，主要归功于两大驱动因素。其一，全社会对AI爆发式激增的算力渴求，远远超出了美国传统“超大规模云巨头”（hyperscalers）新建数据中心算力容量的承接速度。许多新型云厂商脱胎于曾经的加密货币挖矿军团，随后顺理成章地对原有厂房与数据中心进行了改造转型。由于它们摒弃了多元化的复杂通用云业务，将筹码孤注一掷于AI算力这一垂直赛道，因此得以集中所有资金饱和式投入。最终，它们最大的座上宾反而变成了诸如OpenAI这样的顶尖大模型开发商，以及由于自有运力告急而不得不向其包租机架容量的传统云巨头们自己。其二，作为全球绝对垄断的AI芯片巨擘，英伟达急于摆脱对少数几家传统超大规模云巨头的渠道依赖，因而慷慨提供了全方位的金融扶持，全力助推这些新型算力厂商吞下更多英伟达芯片。

CoreWeave、Nebius、IREN、Applied Digital以及Core Scientific这五大上市新型云厂商在过去12个月内的年化营收总和已飙升至180亿美元，而在2024年这一数字仅区区30亿美元。尽管与传统云巨头相比仍有明显体量差距，但它们正在以前所未有的速度蚕食份额。市场研究机构Dell’Oro测算，这些新型云厂商当前合计拥有的数据中心容量已达1.5至2.5吉瓦（GW），而Alphabet、亚马逊、微软和Meta四大巨头的总和约为20至24吉瓦。但新型云厂商的算力规模正以每年翻倍的速度狂飙，扩张步伐几乎是传统云巨头的三倍。

图表来源：《经济学人》

然而，这种令人眩晕的跨越式增长，几乎完全是用债台高筑的庞大杠杆堆砌而成的。这五家上市新型云厂商目前的资产负债表上背负着高达610亿美元的有息负债，仅在短短12个月内就翻了整整两番；此外还有高达180亿美元的经营租赁负债（见图表）。截至目前，没有一家公司能够实现营业利润为正；一旦计入沉重的利息账单，账面更是全线深陷巨额亏损的泥潭。在最近一个季度中，Nebius支付的利息支出相当于其营业收入的五分之一；而CoreWeave的利息吞噬比例更是高达四分之一。

新型云厂商在豪赌一场未来：只要能维持眼下惊人的指数级扩张，盈利的曙光终有一日会降临。或许确实如此。然而，残酷的市场竞争丝毫没有放缓的苗头。除了壁垒森严的传统云巨头——以及正酝酿推出独立第三方云算力服务的Meta——全球目前已有大约200家新兴云厂商在贴身肉搏争抢算力订单。

与此同时，一股不可忽视的中型云计算势力正在加速崛起。原本深耕企业数据库业务的甲骨文（Oracle），为了挤进超大规模云厂商俱乐部而不惜砸下天文数字的真金白银打造专属云体系，其资本支出（大部分靠发债融资）预计在今年将达到770亿美元的高位。今年7月，债台高筑的日本软银集团高调推出了名为SB Neo的全新云服务业务。软银还控股着基础设施企业SB Energy，后者已与OpenAI联手在俄亥俄州投建一座超大型绿色算力中心，并于9月1日正式递交了首次公开募股（IPO）申请。

白热化的同业倾轧，无限放大了新型云厂商头上高悬的两大核心系统性风险。首当其冲的是信贷市场的流动性休克，若融资成本被进一步推高，其本已微薄脆弱的利润空间将面临更具毁灭性的挤压。资本市场上的投资人显然已如惊弓之鸟：新型云厂商发行的企业债收益率与整体垃圾债市场的息差正在急剧拉大。

而更致命、更具破坏性的系统性隐患，在于全产业链供需关系的逆转重塑。眼下，海量算力基础设施之所以能够开足马力大干快上，全维系在一个关键前提之上——即全社会对AI应用的需求将保持超高速的狂飙突进。然而，一旦下游技术的普及放缓，大模型开发商与传统云巨头可能会瞬间斩断外包需求，不再需要这些新型云厂商的高价机位。

上述任何一种极端黑天鹅事件的发生，都将对整个行业带来灭顶之灾。届时，唯有那些除单纯倒卖底层算力之外还能为客户提供增量独特价值的新型云厂商，才可能在这场洗牌中苟延残喘。为此，一部分厂商开始竭力将自己标榜为更值得信赖的基础设施堡垒，以迎合那些对美国科技巨头心存戒备的西方政府；包括CoreWeave和Nebius在内的另一批先行者，则在全力拓展上层软件与配套服务体系。要想在残酷的终局中活下来，新型云厂商必须证明自己绝不仅仅是一个脆弱的二道贩子。■

ch033.xhtml

从蛋白粉到抹茶，食品风潮的蔓延速度前所未见

供应链疲于奔命，难以跟上消费狂潮

乳清需求狂飙 © 盖蒂图片社 (Getty Images)

2026年9月3日

曾几何时，作为奶酪副产品的乳清，其最大的消费者其实是农场里的生猪。如今，连奶牛场农户自己都快用不起了。伴随着全社会疯狂摄入蛋白质的养生狂潮，作为健身奶昔核心原料的浓缩乳清蛋白需求呈现井喷式爆发。然而，乳品加工厂并未相应扩大传统奶酪的生产产能。市场研究机构Vesper的数据显示，在截至今年6月的三年间，乳清蛋白的价格整整暴涨了五倍。同样被TikTok各大网红推崇为“增肌圣品”的茅屋奶酪（cottage cheese），也陷入了如出一辙的严重断货危机。

这两个微观案例生动揭示了整个食品与饮料行业正面临的全新系统性难题：借助社交媒体病毒式裂变，某种小众原料的消费需求能在转瞬之间被推向神坛，然而上游农场与下游工厂的物理产能根本无法在同等速度下实现扩产响应。面对这一现实，企业不得不学会更敏锐地未雨绸缪预测潮流，并在消费者的三分钟热度消退之前顺势套现退场。

食品消费领域追逐潮流的现象其实由来已久。在英国乔治王时代，菠萝曾被奉为彰显尊贵身份的无上象征，以至于催生了专门出租水果供贵族在宴会上摆设充场面的租赁行。到了20世纪六七十年代，瑞士奶酪火锅风靡全美，以至于千家万户购置了数以百万计的专用融酪锅，而今只能被遗忘在橱柜角落里吃灰。然而在今天，社交算法能在短短数周之内，将一股本土风潮无死角卷席至全球各大角落。2023年，一段拍摄于迪拜的开心果夹心巧克力的短视频在TikTok上走红，播放量迅速破亿。短短数月后，各大顶级巧克力制造商纷纷跟风推出自研同款，直接令全球开心果供应链不堪重负。

食品与饮料天生就极易在社交网络上引爆病毒式传播。色泽翠绿的抹茶或是优雅浪漫的香芋紫拿铁，在Instagram上都能拍出极具美感的大片。而且它们单价相对亲民，是普通人唾手可得的精致轻奢。哪怕是尝鲜购买一颗售价高达19美元的日本精品草莓（另一款近期的现象级爆款），也不至于让人囊中羞涩。消费趋势观察博客“新消费者”（The New Consumer）主理人丹·弗罗默指出，几乎将大半生活沉浸在社交媒体上的年轻一代，对一切新奇事物本就抱有更为天然的狂热尝试欲望。

井喷式爆发的畸高需求彻底击穿了上游供应链。2024年至2025年间，伴随着全球对抹茶饮品近乎狂热的追捧，用于研磨抹茶的原料碾茶价格几乎翻了三倍。然而，培育碾茶所需的茶树灌木丛往往需要数年时间方能长成丰产。高烧不退的狂热市场同样引来了各路投机倒把的冒险客。洛杉矶抹茶分销商罗基·徐透露，去年当他库存告罄时，惊觉整个供应链早已被各路蜂拥而至的中间商挤爆。这些新入局者（主要是妄图在种植农户与海外买家之间倒买倒卖抽成谋利的日本投机商人）进一步推高了现货采购成本。

食品企业正绞尽脑汁、八仙过海般应对这一巨变。尽可能提前预判并捕捉风口雏形，使企业能够在竞品反应过来之前提前锁定上游核心原料供应。早在2006年，美国咖啡连锁巨头星巴克便率先推出了抹茶拿铁，远远抢跑在西方席卷而来的抹茶狂潮之前。2024年，在珍珠奶茶彻底成为欧美大众主流文化之前，星巴克便已高瞻远瞩地普及了含有“爆珠”（popping pearls）的特色饮品。最近，它又在旗下的咖啡产品线中全线引入了高蛋白配方，如今其火爆程度在欧洲已与澳白咖啡分庭抗礼。星巴克高管路易丝·迪雷托坦言，公司已经从原先依赖“小范围消费者焦点访谈”的传统模式，彻底转向在全网范围内实时追踪“数以百万计消费者正在即时勾勒的潮流趋势”。她补充道：“真正的核心挑战在于，既要精准判断何时该下场拥抱这股浪潮，又要清醒研判何时入场为时已晚。”

跨界联名合作同样是一剂良方。2022年伴随辛辣风味美食席卷全网，各大快餐连锁巨头排队争抢与当红网红辣油品牌Fly by Jing达成独家联名合作。汉堡连锁巨头Shake Shack更是围绕该品牌定制了一整套限定炸鸡主题菜单。

然而，并非所有的消费狂潮都值得盲目拥抱。星巴克于2024年高调推出的橄榄油风味系列咖啡，在北美与加拿大市场上市不到一年便惨遭全线下架——其背后的致命尴尬在于，大量消费者纷纷投诉饮用后出现了严重的腹泻副作用。■

ch034.xhtml

各行各业的中国企业正在纷纷转向华为

从矿山到交通运输，这家巨头的技术触角正无孔不入

矿山好伙伴 © 盖蒂图片社 (Getty Images)

2026年9月3日 | 海拉尔

如今在中国的各大城市，无人驾驶出租车早已是司空见惯的日常街景。配送无人机同样日益普及。然而，若想亲眼见证人类迄今部署的最重磅自主无人作业机械，你必须一路向北，来到中国极东北部的茫茫大草原深处——伊敏露天煤矿。这是一座横亘在广袤大地之上、占地达42平方公里的超级露天深坑。行至矿坑底部，你会赫然目睹约100辆各载重达90吨煤炭的重型矿卡在井然有序地飞驰——四周却看不到一个人类驾驶员的踪影。原本属于驾驶室的显眼位置，如今全被塞满电池模组的巨大黑色机箱所取代。

负责运营伊敏煤矿的国有电力巨头华能集团业务主管赵耀忠表示，无人驾驶重卡的规模化落地，使得120名一线矿工得以彻底调离采掘坑道内部，转岗至坑外安全的操作岗位——从而将他们从粉尘弥漫且险象环生的井下恶劣环境中彻底解放出来。不仅如此，摆脱了人类生理极限的无人车队工作效率大幅跃升：它们能够不分昼夜、无视极端暴雪天气实现全天候稳定运转，甚至能够自主驶入换电站完成全自动化电池更换。按照既定规划，未来三年内伴随另外300辆全自动矿卡的补充投用，整座露天矿坑底部将基本实现完全的无人化作业。

伊敏露天矿的高效运转，勾勒出中国科技巨头华为正飞速崛起的“企业业务”版图的冰山一角。长期以来，这家巨头以电信设备与消费电子产品闻名于世。2021年，面对核心主营业务遭受美国单边制裁重创的至暗时刻，华为果断宣布其企业业务部门将组建一系列全新的特战编制——即“军团”。这些业务军团旨在打破原有产品线壁垒，整合公司全栈核心技术与工程专长，向机场运营、石油天然气等特定细分企业级行业客户输出一体化赋能方案。

矿山军团正是这种软硬件跨界垂直整合能力的集大成者。漫山遍野耸立在伊敏矿区的华为电信基站，用高密度的5G专网无死角覆盖了整个作业面，为重卡自动驾驶系统的毫秒级响应提供了坚实通信底座。该重卡自动驾驶底层架构脱胎于华为智能汽车解决方案事业部，但针对重型矿卡的庞大车身自重以及矿山崎岖泥泞的极端路况进行了大刀阔斧的底层重构适配。整座煤矿作业调度的统一中控操作系统，正是基于矿鸿（MineHarmonyOS）物联网操作系统构建——而这一系统正是由打造华为智能手机等消费终端的技术团队一手孵化而来。而在许多地下井工矿中，华为云部门开发的大语言及工业视觉AI模型，已被深度运用于地质勘探数据解译与综采工作面设备的远程巡检监测。

作为一家未上市的私营企业，华为并未详尽披露其企业业务军团的细分财务表现。外界最近一次获知其细分营收数据是在2022年，彼时该业务板块已占据其全集团总营收的五分之一，并成为增长最为迅猛的引擎极。华为矿山军团负责人吴安迪表示，该业务板块目前已实现盈利，但由于重工业数字化涉及复杂的定制化工程改造，整体利润率相对微薄。

尽管如此，企业业务对整个华为的战略支柱作用日益举足轻重。8月31日，华为公布的最新财报显示，在2026年上半年，公司净利润同比下滑37%，尽管同期营业收入保持了10%的稳健增长。受全球数据中心建设狂潮驱动，高阶内存存储芯片采购成本急剧飙升，对华为终端消费电子业务的利润空间构成了严峻挤压。与此同时，面对彻底失去获取包括尖端芯片在内的美国核心技术供应的残酷现实，华为不得不将巨量真金白银反哺于核心底层根技术的全栈研发——仅在今年前六个月，高昂的研发开支就一口气吞噬了总营收的整整四分之一。

由于西方政府普遍捕风捉影地担忧华为设备可能被用于网络间谍活动并借此施加全方位准入禁令，在过去十年间，华为的海外市场份额遭到了严重侵蚀。这使得在其本土大本营深耕开拓全新增长引擎变得前所未有地紧迫，尤其是在中国全国范围内的5G宏基站铺设已基本宣告收官的大背景下。对于华为而言，幸运的是，中国政府正在从制造业、综合交通运输乃至智慧电网调度等方方面面，以举国之力全方位推动人工智能在实体经济全产业链中的规模化渗透落地。华为麾下枕戈待旦的企业业务军团，已然做好了全面出击的准备。■

ch035.xhtml

美国国税局正全面追剿“美国企业界”的海外利润

从可口可乐到Meta，跨国巨头无一例外被列入猎杀名单

图片来源：盖蒂图片社 (Getty Images)

2026年9月3日 | 纽约

几十年来，美国跨国企业巨头一直深谙此道：通过将巨额账面利润移花接木转移至海外避税天堂，神不知鬼不觉地粉饰财报净利润。据倡导税收透明的权益组织FACT联盟测算，仅在2025年一年，美国企业凭借此类资本腾挪财技至少变相截留了110亿美元的税款。然而如今，联邦政府终于准备动刀收回这笔横财了。

去年，当唐纳德·特朗普重返白宫椭圆形办公室的首日便高调宣布废除前任政府主导的多边全球最低企业税率框架时，无数美国大公司掌门人曾如释重负。他们曾天真地以为，这意味着跨国巨头将免受政府对跨国避税行径的补充追缴。然而，那些寄望于联邦政府自此对海外利润网开一面的乐观派，显然高兴得太早了。今年5月，美国健康保险巨头联合健康（UnitedHealth）主动披露，美国国税局（IRS）正因其将大笔利润人为归入海外离岸子公司展开强力税务稽查，税务部门强硬认定该公司在本土存在严重的偷逃税款行为。此案标志着国税局自奥巴马执政时期启动的针对所谓“转让定价”（transfer pricing）灰色操作的强力清算行动正在全面铺开——即便在奉行保护主义的特朗普任内，这一追缴势头也丝毫未见任何松动回撤的迹象。涉及其中的补缴金额高达数百亿美元。

在这一轮因转让定价涉税纠纷被国税局死死咬住的企业名单上，生物医药先锋安进（Amgen）以及旗下坐拥Facebook和Instagram的社交媒体霸主Meta赫然在列。然而，在所有被调查对象中，针对全球饮料巨擘可口可乐的税案，被税务司法界公认为“转让定价争议领域的超级碗赛事”（足见这群税务律师平日的自娱自乐有多么乏味）。

这场已在司法程序中拉锯缠斗了逾十年的超级诉讼，有望在未来数月内迎来终审裁决。早在2015年，美国国税局便向可口可乐开出巨额补税罚单，指控其在2007年至2009年纳税年度间少缴了33亿美元税款，直指该公司违规将源自核心知识产权（包括其全球闻名的经典商标权）所产生的巨额利润人为截留在海外子公司名下。由于这场法律拉锯战在法院系统中层层审理历经数年，可口可乐在此期间依然有恃无恐地沿用同一套避税会计策略。其直接后果是，该案涉及的潜在补税及罚款总额已如滚雪球般膨胀至惊人的200亿美元之巨。可口可乐辩称，其海外子公司在当地市场投入了极其关键的营销与品牌建设资本，理应匹配其所录得的天文数字般的资本回报率；该公司同时坚称，自己仅仅是在严格恪守国税局早在1996年与该公司签署的一份关于跨国利润分配的先期定价谅解备忘录（尽管国税局认定该协议仅适用于历史收益）。

面对国税局在转让定价问题上的步步紧逼，可口可乐及其他涉案跨国巨头在公开场合却表现得异乎寻常的云淡风轻。截至今年7月，这家饮料巨头账面上仅预提了区区5.3亿美元的准备金，来应对一旦最终败诉可能面临的高达140亿美元的天价净应补税单（此前公司已依法暂扣补缴了60亿美元，若最终胜诉该笔资金将被退还）。今年7月，安进公司刚刚就一起投资者发起的集体诉讼达成庭外和解，原告指控其向资本市场隐瞒低估了其与国税局涉税诉讼的系统性风险。在此案审理期间，一名联邦主审法官曾一针见血地辛辣嘲讽道：安进向投资者的风险披露方式，“就像一个小孩在偷吃完整整一大块蛋糕后，若无其事地告诉父母自己只是吃了一点‘饭后甜点’”。

全球知名会计师事务所致同（Grant Thornton）的史蒂文·拉普指出，倘若国税局能够最终打赢手中这批体量最大的转让定价官司，连本带利外加行政罚金，将能从少数跨国巨头手中一次性追缴回大约1000亿美元的巨额财税收入。这一规模几乎相当于美国联邦政府在2025年征收的企业所得税总额的五分之一。

短期来看，想要进一步全面扩大这一税收追剿风暴面临着不小的现实阻力：在特朗普大刀阔斧削减联邦开支的大旗下，美国国税局的编制与预算遭遇了大幅裁撤，一线稽查人员已然捉襟见肘、疲于奔命。然而从长期趋势审视，跨国转让定价纠纷的井喷已成不可逆转的历史必然。其核心原因之一在于，现代跨国企业的核心资产价值中，越来越多的比例是由无形资产构成的——这类资产无需进行任何实体跨国转运，为跨国避税账面腾挪提供了得天独厚的绝佳载体。知识产权资产评估机构Ocean Tomo的数据显示，从1975年到2025年的半个世纪里，标普500指数成分股公司的总企业价值中，无形资产的占比从最初的17%一路飙升至令人咋舌的92%。另一重无法忽视的残酷现实则是——正如债市投资者心知肚明的那样——美国联邦政府债务规模正以脱轨之势疯狂狂飙。大洋彼岸庞大的跨国寡头们，恐怕迟早要为这笔天文数字般的国家账单分摊更大份额的真金白银。■

ch036.xhtml

LIV高尔夫联赛的溃败，给体育商业投资上了生动一课

颠覆一项传统体育运动，绝非挥金如土就能一蹴而就

图片来源：盖蒂图片社 (Getty Images)

2026年9月3日 | 洛杉矶

抛开那些令人咋舌的天价球员工资、略显荒诞的“让高尔夫更喧嚣”的品牌口号，以及球道上震耳欲聋的流行电音背景乐不谈，LIV高尔夫联赛（LIV Golf）从创立伊始，骨子里始终带着一丝不知天高地厚的狂妄蛮劲。沙特阿拉伯，一个地形地貌酷似沙坑、国内毫无高尔夫历史底蕴的中东国度，竟妄图以一己之力颠覆全球最具门槛、最为贵族清高的传统体育运动。不出所料，这场鲁莽的跨界颠覆最终以惨烈败局草草收场。上个月，规模达9000亿美元的沙特主权财富基金——公共投资基金（PIF），正式宣布彻底切断对其输血。在接下来的数日内，LIV高尔夫联赛预计将正式向法院递交破产申请。

然而，溃败的领域远不止绿茵果岭。在过去的五年时间里，PIF主导了沙特在体坛掀起的一轮横扫全球的超级投资狂潮，触角全面延伸至足球、网球、职业拳击、综合格斗、赛车运动以及新兴电子竞技。其狂野的扩张雄心令人侧目。彼时国际社会铺天盖地的批评声浪，指责沙特王储穆罕默德·本·萨勒曼试图用滚滚石油美元砸向职业运动员的腰包，以此大搞“体育洗白”（sportswashing），掩盖该国专制残暴的国际人权形象。然而这种批评至少在很大程度上未切中要害。PIF起初的底层愿景，是寄望利用高对抗体育赛事的感召力，让本国长期处于亚健康状态的年轻一代动起来；同时通过在本土承办全球顶级殿堂级体育赛事招揽八方游客，从而加速其高度依赖原油的经济结构转型。

然而时至今日，传统实体竞技体育却被无情打入了冷宫的保级泥潭。在PIF最新公布的《2026-2030年战略规划纲要》中，长达50余页的报告里，“体育”（sports）一词仅仅被提及了区区七次——且其中六次的前面都赫然带有“电子”（e-）前缀。正当大批私募股权基金与跨国机构投资者如飞蛾扑火般将数以百亿计的资本注入全球体育俱乐部特许经营权、以至于不少人惊呼出现严重泡沫之际，沙特资本的全线败退能给市场带来哪些深刻的警示与镜鉴？

首当其冲的第一要义，是永远不要低估体育迷的情感羁绊。坐拥取之不尽用之不竭的国库支票簿是一回事，但要想用金钱彻底买断并俘获球迷的芳心，则是完全不同的另一回事。PIF为了扶植一个能够与美国美巡赛（PGA Tour）分庭抗礼的全新联盟，在LIV高尔夫项目上肆意挥霍了整整50亿美元，试图凭借挥金如土的巨额奖金池、节奏更快的分站赛制以及令人眼花缭乱的声光电秀来吸引年轻受众。然而，尽管重金挖角了一批顶尖明星球员，LIV那套浮躁奢靡的作派却彻底惹恼并劝退了高尔夫领域最核心的高净值传统拥趸——这群人最痴迷的恰恰是PGA数百年沉淀下来的绅士传统、历史厚重感与崇高声望。由于核心收视基本盘寥寥无几，LIV根本无法通过转播权商业变现形成自我造血的正向循环。

核心球迷往往同样是一股足以颠覆商业规划的毁灭性力量。国际足联（FIFA）前不久试图通过向外资拆分出售其皇冠上的明珠——包括男足世界杯特许经营权在内的核心商业资产，以此变相融资42亿美元，该激进举措最终被迫紧急叫停并全面搁浅。这一惨败正是由于遭到了全球各国足协的强烈反弹，而各国足协最大的顾虑，莫过于彻底引爆数以亿计狂热足球铁杆拥趸的无边怒火。

由此引申出第二重教训：在体育产业中，所谓的“创造性破坏”举步维艰。正如对待诸多历经数代人文化积淀的百年事业一样，外来资本极易产生一种傲慢错觉，误以为传统体育是一片亟待通过互联网资本逻辑颠覆的熟透果园。除PIF之外，不少雄心勃勃的投资机构也曾试图通过创办炫酷的新型衍生联赛来打破垄断格局，最终却无一例外折戟沉沙。由美国传奇短跑名将迈克尔·约翰逊牵头成立、旨在正面对标田径钻石联赛的“大满贯田径赛”（Grand Slam Track），在首站鸣枪后不到一年，便在去年12月匆匆宣告破产清算。只有极少数另辟蹊径的后来者得以突围，例如美国终极格斗冠军赛（UFC），这家综合格斗巨头将血腥的八角笼搏击成功运作推向了前所未有的商业高度。然而UFC的成功密码在于凭空孵化出了一个全新的增量市场，而非试图硬闯那些早已被百年传统豪强死死盘踞的成熟红海。

第三重残酷教训在于“时间”——或者说“时机”。打造一个经久不衰的王牌体育IP，往往需要耗费数代人的漫长心血。自巴西球王贝利加盟纽约宇宙队、开启北美足球启蒙至今，已经跨越了半个多世纪的风雨沧桑。直到今天，美国职业足球大联盟才勉强真正形成气候与商业价值（而当年贝利效力的宇宙队早已尘封在历史的故纸堆中）。在当今资本狂欢的当下，赚快钱最立竿见影的法门，或许莫过于高位接盘一支历史底蕴深厚的知名豪门，随后迅速转手倒卖套现。金融巨鳄马克·沃尔特在上个月正是施展了这一财技：他以创纪录的125亿美元天价，将旗下洛杉矶湖人队的大股东权益成功倒手甩卖给两位超级富豪同行，短短一年间便净赚了惊人的25亿美元溢价。然而在极度亢奋且充满流动性泡沫的市场中，这种击鼓传花的高频倒买倒卖无疑是在刀尖上起舞。

倘若传统实体路径全面受阻，那就必须将战略纵深投向更加长远的未来。沙特阿拉伯在跨界投资那些横跨数百年的传统老牌体育运动时，战绩堪称一地鸡毛、惨不忍睹。但在过去五年间，它在布局方兴未艾的电竞领域时却展现出了异乎寻常的条理与章法，其中包括在上个月刚刚以550亿美元巨资完成对全球游戏巨头艺电（Electronic Arts）的私有化世纪并购。坐在屏幕前收看电竞全球总决赛或许无法让沙特年轻人告别沙发走向户外锻炼——但他们发自内心地为此癫狂。从资本回报与战略协同的角度审视，这笔布局显然远比当年堂吉诃德式强推LIV高尔夫要清醒现实得多。■

ch037.xhtml

企业内部培训视频的“绝妙荣光”

简笔动画、奖励小红星外加浮夸演技：经典要素一个都不少

插画摄影：保罗·布洛 (Paul Blow)

2026年9月3日

职场技能培训有时确实能传授一些立竿见影的硬核技能，它有助于规避潜在的法律合规诉讼，完成各项强制性的行政合规指标。与此同时，它还是体验时空穿越最廉价便捷的法门。你只需耐着性子点开一部标准的企业合规培训视频，便能瞬间产生一种时光倒流几十年的恍惚错觉。

首先，也是最显而易见的一点在于，其中不少视频的拍摄年代可能比你的实际年龄还要古老，其探讨的主题往往“常青”到几十年来根本无需任何版本迭代。画面粗糙模糊、雪花闪烁。镜头里的男士一律系着呆板的窄版领带，女士则无一例外穿着百褶裙，彼此呼唤着诸如“罗德尼”和“卡罗尔”之类极具年代感的名字。没有一个人身上有纹身，更没有人会在工位上展现个性化的真实自我。在那个年代的剧本里，员工如果告病请假，绝不可能是什么心理健康问题或焦虑症，而一定是某种早已绝迹的古老恶疾，比如肺结核或瘰疬。

当然，有些培训视频或许是近几年才刚刚杀青的，但成片效果依然透着一股难以名状的土味与过时感。这主要是由于这类片子制作预算极其逼仄廉价，请来的尽是些接不到正经通告的三流群演；为了尽可能延长视频的使用寿命，导演往往被严格授意：画面中绝不能出现任何带有明显时代特征的前沿现代元素。

举例而言，随便点开一部关于职场反骚扰的合规警示短片，你大概率会看到几个身着荧光警示反光背心的男人在工地上假模假样地面红耳赤互相推搡咆哮，而画外音则语重心长地警告称：同事间的“打闹嬉戏”（horseplay）极易擦枪走火失去控制。打闹嬉戏？这个词到底是从哪个中世纪古董堆里扒出来的？随后旁白谆谆告诫大家：职场冷暴力骚扰会让员工陷入深深的孤独与排挤感——镜头应声切换到一名孤苦伶仃坐在挖掘机驾驶室里的工友脸上，后者极力挤出一副写满被全世界抛弃的茫然表情。相信我，你这辈子上一次看到这么生硬拙劣的演技，恐怕还是在幼儿园的圣诞迎新话剧汇演上。

第二重灾难在于，海量的企业培训材料所聚焦的议题，往往弱智到根本无需任何形式的人类智力培训。坐在屏幕前观看某些企业培训视频，你会被迫接受一场被当成低智儿童对待的人格洗礼——更确切地说，其设定仿佛是默认屏幕前的观众，连一个低智儿童眼里的傻瓜都不如。你的专栏作家最近刚刚硬着头皮刷完了一套关于“防范工作场所滑倒、绊倒与跌落”的强制在线通识课，全程看完后，内心不禁陷入对人类这个物种究竟是如何奇迹般繁衍存续至今的深度哲学怀疑。

课程中充斥着对湿滑落叶以及从室外水泥地跨入室内光滑瓷砖地面潜在致命风险的反复警示，期间甚至煞有介事地留出了若干分钟的“沉浸式深度反思时间”，要求学员闭门思索如下灵魂叩问：“在你的日常工作与生活中，最常导致你滑倒或绊倒的障碍物究竟有哪些？”不仅如此，片中还语重心长地抛出了一系列颠覆认知的新奇洞见：事实证明，要有效防范雨天滑倒，只需在跨入大楼正门时，主动在大厅的地垫上擦拭干净你的“鞋底”。请务必划重点注意“鞋底”这个用词——毕竟在编剧看来，一个未经培训的门外汉员工最明显的特征，就是在试图擦干鞋子时，会整个人毫无尊严地仰面平躺在大理石地面上，像一头处于发情期的象海豹一样手舞足蹈、胡乱扑腾着去蹭鞋面。

第三，这类培训视频极度依赖并照搬了全套小学低年级的应试驯化套路。在那些面部僵硬的三流演员退场后，屏幕上便会无缝衔接粗糙低幼的Flash简笔动画。每隔三五分钟，系统便会强制弹出一道单项选择题，以此死板检验你的理解吸收能力。在那门经典的防滑倒防绊倒课程中，剧情安排了一名端盘子的餐厅侍应生，因未能及时察觉“用餐区入口处的装饰地毯防滑衬底已老化失去抓地力”，导致脚尖不慎被卷起的毯边挂到，身形一个趔趄后踉跄稳住了重心。系统随即严肃发问：请问该员工刚才经历的属于以下哪种突发意外事件——A. 滑倒，B. 绊倒，还是 C. 跌落摔倒？

很难想象还有比一名餐厅服务员差点摔倒却又没真正摔倒更无聊琐碎的模拟场景了，但系统却极其擅长拿捏你内心深处尚未泯灭的幼童心性。在培训视频里只要答对一道选择题，界面就会“叮”的一声赏你一枚虚拟数字徽章或一颗闪闪发光的小红星。一旦你满分通关所有测验，整个屏幕便会瞬间绽放出漫天飞舞的虚拟彩色碎纸屑特效，系统还会贴心地为你生成一张支持一键下载打印的个人成就荣誉证书。反之，倘若未能在截止日期前刷完所有课时进度，你的人事黑名单便会被一键通报给你的部门直属主管。

当你把一个成年群体彻底当成幼儿园孩童来规训对待时，他们的行为反弹往往会以惊人的速度退化成儿童般的幼稚法则。哪怕你是一位即将光荣退休的资深职场老将，在集齐一颗系统奖励的小红星时，内心深处依旧会莫名泛起一丝隐秘的爽感与成就感。你甚至会莫名其妙地对自己刚刚斩获的那张“防滑倒、绊倒与跌落初级认证证书”引以为傲，郑重其事地将其打印装裱后用图钉按在逼仄的办公隔间挡板上。夜深人静看着证书时，你甚至会暗自琢磨自己要不要顺手把那门“如何安全使用人字梯”的高阶微课也给考下来，转念一想又不禁对自己这股突如其来的疯狂野心哑然失笑。

与此同时，你自然也会毫无心理负担地向内心的作弊本能彻底缴械投降。在点开一部关于“如何防范钓鱼邮件、短信钓鱼与二维码伪造投毒”的强制网络安全通识课时，你顺手将视频窗口静音并缩至屏幕角落，任由其后台无声播放，只有在进度条卡住必须手动点击“下一页”时才象征性抬眼看上一眼。一旦发现答错多选题根本没有任何实质惩戒措施，你甚至会毫不犹豫地跳过通篇课件直奔文末自测题，像玩扫雷游戏一样闭着眼睛一顿乱点，直到瞎猫碰死耗子刷出全对的优异成绩，顺利领到属于你的钓鱼防护认证通关证书。

这一切与当年的应试校园别无二致：你手里攥紧了一纸光鲜亮丽的合规资质证书，但脑子里却压根没有真正吸收任何有价值的常识。正因如此，当你在那天傍晚收拾公文包跨出办公大楼的玻璃旋转门时，你压根没有留心注意到——某位童心未泯的恶作剧同事早已在大厅的玄关瓷砖地面上密密麻麻铺满了一层湿漉漉的烂树叶，仅仅是为了在下班前搞一场毫无恶意的“打闹嬉戏”。■

ch038.xhtml

能看体育纪录片，何必去读 MBA？

商学院的管理灵感已近枯竭，更衣室里却俯拾皆是

插图：布雷特·莱德

2026年9月2日

长期以来，全球顶尖商学院一直在放低身段，把进修证书卖给任何拥有电脑和闲钱的人。如今，买单者甚至只需支付 699 美元，就能接受哈佛商学院人工智能生成教授的指导。然而，在亚马逊学院和奈飞大学的体育分院里，还有一门更便宜（且同样声名显赫）的商业速成课正虚席以待。

体育纪录片通常不被视作管理学的经典正典。20 世纪 90 年代中期，著名商业思想家克莱顿·克里斯坦森正在撰写关于“颠覆性创新”的著作，而英国电视台则制作了《格雷厄姆·泰勒：一项不可能完成的任务》，对时任英格兰国家足球队主帅进行了略显灰暗的白描。2005 年，两位欧洲工商管理学院（INSEAD）的教授在一本畅销书中告诫老板们去开拓“蓝海”市场。而另一位英格兰足球主教练、纪录片《沃诺克》中克己自律的主角尼尔·沃诺克，在赛前对球员们这样喊道：“尽情享受比赛，但必须以他妈的纪律为前提来享受。”

如今，大概念枯竭的换成了商学院，更衣室里反倒涌现出源源不断的新思潮。流媒体平台意识到观众对场边哲学的渴望近乎无穷无尽——这始于疫情期间大获成功的奈飞赛车纪录片《极速求生》（记录一级方程式车队，如今已播至第九季），以及讲述芝加哥公牛队 1997-98 赛季篮球征程的十集纪录片《最后之舞》。

眼下体育纪录片泛滥成灾，单凭今年仅在奈飞上线、且仅聚焦男子足球的作品，就足以攒出一整门 MBA 课程。想要学习企业逆风翻盘的老板，可以先看看讲述利物浦在 2005 年欧冠决赛上演惊天大逆转的片子，该片片长 78 分钟，远长于那场仅耗时约六分钟的翻盘本身。正在权衡拓展新市场的高管，可以参考跟拍挪威国家队的纪录片《挪威：黑马》，快速入门斯堪的纳维亚式管理。而《大巴车：法国足球内讧纪实》，则是一部教科书级的职场毒性文化反面教材。

这类影片大多充斥着浮夸吹嘘，缺乏实操指导。不过，亚马逊近期推出的《美丽的执念》堪称不可多得的管理学明珠，该片跟拍了佩普·瓜迪奥拉执掌英超豪门曼城队的最后一个赛季。瓜迪奥拉执掌着庞大的业务体量（该俱乐部年营收约 10 亿美元），其工作绩效每周都要接受考核。作为兼具本土底色与全球声誉的品牌掌舵人，瓜迪奥拉不仅要向绿洲乐队的诺埃尔·加拉格尔交代，还要向阿布扎比主权财富基金的哈勒敦·阿尔·穆巴拉克述职，后者的董事长正是俱乐部老板。

多数管理潮流往往依托于单一且鲜明的主张，瓜迪奥拉则是一位极具分裂风格且极其成功的管理者。在某场比赛前，他执意要求更衣室里人人拥抱；而在输掉比赛后，气氛急转直下：“谁要是敢对我、对我的决定摆出一副抱怨的臭脸，我他妈弄死你，”他怒吼道。队内罕有的传统管理话术源泉是队长凯尔·沃克，他总能吐出诸如“场上一定要团结一心”之类的金句。但在第一集结尾，沃克就遭到了扫地出门的对待。

商学院里看似新颖的理论，在球场边不过是寻常的常识。去年发表的一篇论文中，研究人员利用“基于神经网络的机器学习”证明，在艰难时期管理公司所承受的压力会加速老板的衰老。瓜迪奥拉在承受巨大压力时会用指甲狠抓头皮，抓出的道道红印恰恰印证了这一观点。他的方法论甚至带有不少“数字化转型”的痕迹——一套涵盖 43 项不同指标的模型为球员评分并辅助人员任用决策——尽管职业体育界依旧是白板战术推演的最后堡垒。

多数老板大概会觉得瓜迪奥拉的更衣室严酷且陌生。他们理应扪心自问为何会有这种感觉。原因之一在于那里见不到女性的身影。但最主要的原因，是寻常人对胜利的渴望远不如瓜迪奥拉那般狂热。这无可厚非：如果一家公司的中层管理人员在推销汽车配件等业务时也照搬这种风格，那这家公司将成为极其可怕的人间地狱。然而，当今经济很大程度上是由人工智能实验室和对冲基金这类组织驱动的，它们极度依赖将极少数顶尖奇才的罕见天赋压榨到极致。外界极难窥探此类机构的内部运作，而这正是瓜迪奥拉的执教纪录片极具观赏价值的原因所在。

眼下投资者正以创纪录的速度豪掷重金收购体育俱乐部特许经营权，他们有更务实的理由去恶补体育纪录片。当沙特阿拉伯仍在为 LIV 高尔夫巡回赛的受挫而舔舐伤口之际，或许可以通过观看福克斯体育频道去年播出的相关系列纪录片来做个复盘。如果乔舒亚·库什纳早先看过了那些展现球迷对足球是何等痴迷的纪录片，或许就不会去尝试推动那场以流产告终的世界足球私有化计划了。

绿茵场上的工商管理硕士

体育队伍是极其特殊的机构。对于坐拥数十亿资产的富豪而言，它们固然是终极的门面资产；但同时也是商业变现的载体，且许多人坚信体育基本上不受人工智能颠覆浪潮的影响。狭义的论据在于，与电影或电视节目不同，现场体育赛事无法由人工智能模型凭空创造。广义的论调则认为，我们正在迈向一个失业人群需要打发闲暇时光的世界，他们的绝大部分注意力（或许还有身份认同）都将寄托在体育队伍身上。在这样一个世界里，球队主帅将摇身一变成为社会首屈一指的哲学家，而奈飞那些记录幕后细节的纪录片，则将成为他们的传世巨著。■

ch039.xhtml

迈阿密经济的非凡崛起

借力资本与政治的双重助推，这座城市繁荣正盛。但它的成功可以持续吗？

摄影：盖蒂图片社

2026年8月31日 | 迈阿密

在迈阿密的一个周五夜晚，一位豪宅中介刚走出一家高档酒廊，伫立在一座摩天大楼外。他正在庆祝自己刚卖出一套豪宅——售价高达 5000 万美元。他虽不愿透露买家姓名，但表示自己的客户多为科技巨头和金融大鳄。至于他身后的摩天大楼“布里克尔 830 号”，底部几层略显陈旧，楼内却拥有全美最为昂贵的写字楼空间。大型对冲基金城堡投资、科技亿万富豪彼得·蒂尔家族办公室名下的一家投资公司，以及两家顶尖律所均坐落于此。

图表：经济学人

这位房产中介和布里克尔 830 号正是迈阿密崛起的缩影。从 2020 年到 2024 年（有数据可查的最近年份），该大都市圈的国内生产总值（GDP）增长了 23%，而全美同期增速仅为 13%（见图表 1）。各路富豪纷至沓来。新定居者包括城堡投资掌门人肯·格里芬、亚马逊及太空公司蓝色起源（在佛罗里达州制造并组装发射火箭）的创始人杰夫·贝佐斯，以及 Meta 掌门人马克·扎克伯格。其背后不仅有金融资本的注入，更有政治力量的烘托：唐纳德·特朗普除了华盛顿特区外，在此处周边停留的时间超过了其他任何地方，且大多待在他的海湖庄园。棕榈滩已然成为政策制定者与外交官云集的枢纽。

“十五年前，”这位房产中介说，“唯一觉得在这里当个富人很明智的，只有古巴人、DJ 和高尔夫球手。”话虽有些夸大：迈阿密长期以来都是区域中心。但其昔日的经济确实高度依赖毗邻拉丁美洲的区位优势，以及有时略显浮夸的娱乐产业。当时迈阿密市中心的经济规模仅约为曼哈顿的四分之一。

如今这一比例已达二分之一，且仍在上升。2020 年，迈阿密没有一栋住宅售价能达到 5000 万美元。到了 2025 年，这一数字飙升至 17 栋，几乎占到了全美该价位住宅成交总数的一半。许多举家迁入或将业务迁来的亿万富豪与企业坚称，之所以选择这里，是因为他们坚信该市的经济规模很快将跻身全美前列，足以比肩纽约、芝加哥与旧金山。迈阿密的这波繁荣能否持久尚存争议，但就眼下而言，它正乘着金融资本与政治权力深度交融的强劲东风茁壮成长。

过去五年间，三股热潮先后席卷南佛罗里达海岸。第一股热潮带来了海外资金。2025 年前一直担任迈阿密市长的共和党人弗朗西斯·苏亚雷斯曾周游世界招商引资，在波斯湾地区成效尤为显著。如今当地各大酒店挤满了来自中东的对冲基金经理、交易员和顾问。自特朗普对伊朗开战以来，部分人士甚至选择永久定居于此。沙特阿拉伯在当地设立了办事处；海湾国家主权基金从酒店到加密货币交易所均进行了广泛投资。此外，自拉美部分右翼政府放宽资本管制后，涌入的拉美投资者数量进一步激增。

规模更为庞大的一股热潮则来自美国本土：受够了其他城市高昂税负和左翼政策的富人与企业纷纷南迁。佛罗里达州不征收个人所得税；而纽约市富裕阶层的综合税率高达 14.8%。在佛州，顶级豪宅的房产税甚至不足 1%。长期以来，美国富人一直习惯去佛州度假休闲，流连于 24 小时通宵营业的夜店或是高端养老居所。然而近期，许多人直接在此安家落户。美国国税局的数据显示，2025 年迁入该州的人群年收入总和达到了约 400 亿美元。

在企业层面，佛罗里达州各级政府出台了全方位的税收减免与激励政策，例如向给员工发放高于全州平均工资待遇的企业提供税费减免。迈阿密官员更是展现出坚定的亲商姿态。当格里芬开始抱怨城堡投资此前总部所在地芝加哥的治安与税负问题时，苏亚雷斯便主动致电，力邀其南迁。目前已有至少 150 家金融机构完成迁移，包括银行业巨头富国银行的私人财富管理部门，以及埃利奥特管理公司等城堡投资的业内同行。科技企业同样阵容强大。拥有大量政府订单但颇具争议的人工智能数据分析巨头帕兰提尔（Palantir）于 2026 年从政治立场偏左的丹佛迁出。来自旧金山的落户企业中，就包括蒂尔旗下的三家公司。

第三股热潮则由南佛罗里达最著名的富豪移民领衔。特朗普将这里打造成了一个政治神经中枢。在去年 10 月至今年 3 月期间，他有四分之一的时间在此度过。多名佛罗里达政界人士进入了他的核心权力圈，包括白宫办公厅主任苏茜·威尔斯、司法部长托德·布兰奇以及国务卿马尔科·卢比奥。威尔斯的一名下属透露，特朗普在第一任期内就曾频繁前往海湖庄园，“但这一次，我们更多关键的实质决策都是直接在佛罗里达敲定的。”

在海湖庄园，特朗普接待了多国政要，并敲定了对古巴实施石油封锁以及抓捕委内瑞拉独裁者尼古拉斯·马杜罗的计划。春季联邦预算出台前，他的幕僚在此与企业首席执行官举行会晤；7 月总统签署修正“联邦医疗保险”（针对老年人的医疗计划）前，也在庄园听取了医疗机构代表的汇报。“听着，”一位官员表示，“在总统看来，全美最聪明的一批人本就早是海湖庄园的会员了。”

随着政治活动重心的南移，资本亦步亦趋。全美最大的游说机构巴拉德合伙公司（发源于佛罗里达）如今在棕榈滩设立了其规模最大的分支机构。多国政府纷纷增派派驻外交官。“我们的边防部门和国防部，”一位总领事直言，“都必须能够随时接洽到棕榈滩。”三个欧洲国家的外交官员透露，其驻迈阿密领事馆的人员规模已然超过了驻纽约领事馆。

图表：经济学人

在这一背景下，房地产业的蓬勃发展便不足为奇。2025 年，房地产占迈阿密经济比重达 18%，高于 2014 年的 14%。自疫情暴发以来，建筑业就业岗位增长了 12%（见图表 2）。格里芬正斥资约 20 亿美元在此兴建一座楼高 54 层的全新摩天大楼。

金融业同样发展亮眼：调研数据表明，得益于华尔街人才的大批迁入，自 2019 年以来后台运营岗位激增了 20%。法律咨询等企业专业服务在当地 GDP 中的比重上升了两个百分点。

作为传统经济支柱的接待业依旧红火。该行业吸纳了迈阿密都市圈十分之一的就业人口，这一比例在全美名列前茅。迈阿密是全美唯一一座同时拥有北美四大体育联盟球队、顶级赛车赛事以及顶级网球赛事（迈阿密公开赛）的城市。在特朗普 2025 年出席的四场商业会议演讲中，有两场选在了迈阿密。“我们一直充满魅力，”奢华酒店法埃纳的一位娱乐经纪人表示，“但这些新来者的涌入为我们赋予了专属的高端稀缺性。”

迈阿密的这场繁荣究竟会演化为稳健持久的扩张，还是终将戛然而止？成功的都市离不开强大的网络效应：无论金融、科技还是公关游说，企业总是流向确信能够招募到客户与专业人才的沃土；人才也会奔向能满足职业抱负的集聚区。要实现这一跨越，迈阿密必须跨越三道路障。

图表：经济学人

第一大阻碍在于潜在的人口分化——缺乏非富裕阶层人口。财富的集中涌入将当地生活成本推升到了令人咋舌的高度。根据美国商务部经济分析局的数据，迈阿密的住宅租金已高出全美平均水平，甚至反超了纽约。自 2020 年初以来，当地消费者物价指数飙升了近 40%，而全美平均涨幅为 28%（见图表 3）。随着奢华餐饮网点遍地开花，平价商铺被逐步挤出。高昂的生活开销正迫使底层贫困人口加速撤离。自 2014 年以来，该市百万富翁数量几近翻倍；但在 2024 年 7 月至 2025 年 7 月期间，迈阿密总人口却缩水了 1 万人。为接待服务业输送基层劳劳动力的贫民街区，人口流失速度最为迅猛。

第二大短板在于智力资本与科研储备的不足。纽约、西雅图和旧金山属于典型的知识型经济体，源源不断地吸引着高端技术人才以及驱动这些人才的资本。尽管近期吸纳了不少顶尖才俊与财富，迈阿密与这些竞争对手相比仍差距悬殊。作为美国第四大经济强州，佛罗里达州内的企业与大学在 2023 年仅投入了 130 亿美元用于研发，在前十大经济强州中垫底。佛州拥有大学文凭的人口比例亦低于全美平均水准。迈阿密那些新晋腰缠万贯的风险投资家所配置的大部分资金，最终依然投向了旧金山的初创公司。

格里芬曾承诺，到 2031 年，城堡投资旗下将有更多员工迁至迈阿密办公。目前常驻迈阿密的员工已有 400 人，但其中绝大多数并不从事核心的投资顾问工作。如果剥离后台职能岗位，金融业占迈阿密经济总量的比重其实并不比十年前高。作为迈阿密规模最大的白领产业，当地金融业的体量仅为纽约的八分之一。

第三重隐忧在于气候环境。南佛罗里达位于近海平面的半岛之上，极易遭受极端天气的侵袭。气象预报机构 AccuWeather 的一名气象学家直言，全美没有哪座城市比迈阿密更容易遭到飓风摧毁。今年 8 月，发表在《科学报告》上的一篇最新学术论文预测，若全球气温较工业化前升高超过 1.5 摄氏度，风暴潮将导致迈阿密频繁被洪水淹没。目前，佛罗里达州已成为全美按揭购房房屋财产保险费用最昂贵的州。

即便如此，迈阿密眼下的发展势头依然不容置疑。当地一位态度乐观的政界人士坚信，就像爆发伊战前的波斯湾一样，这里可以成为全球商界拓展版图、储存财富的理想之选。一年前，特朗普在海湖庄园设宴款待了 33 位美国商界领袖。到场嘉宾中有半数以上（包括特朗普、威尔斯及其配偶）都在佛州购置了房产。2029 年特朗普任期届满之后，佛罗里达的政治吸附力可能会逐渐消退。或许这座城市凭借持续吸引富豪就能维持繁荣；但或许，它所需要的远不止这些。■

ch040.xhtml

欧洲债券市场遭遇“节后综合征”

收益率飙升的诱因虽与美国有所不同，但棘手程度分毫不减

摄影：马蒂亚斯·巴苏阿尔多/EPA/Shutterstock

2026年9月1日

对欧元区主权债券感兴趣的投资者正面临着特殊的挑战。除了欧盟自身发行的债券外，他们还有 21 个国家的国债可供挑选。夏季的漫长休假（包括交易员暂离交易台）通常意味着市场在此期间波动剧烈。此外，由于某些神秘规律，作为欧洲基准的德国十年期国债收益率往往在 8 月过后攀升 0.15 个百分点。

图表：经济学人

今年，这一异动提前到来，且波及范围远不止德国。近几周来，全球债券市场情绪躁动不安（见图表 1）。因通胀风险上升、经济增长前景不明朗以及各国政府毫无紧缩开支的意愿，心惊胆战的投资者纷纷抛售主权债券（债券价格随收益率攀升而下跌），导致全球各地的国债收益率几乎全线上扬。9 月 1 日，日本十年期国债收益率三十年来首次触及 3%。美国 30 年期国债收益率已十分逼近 8 月中旬创下的 5.31% 的 19 年来高位。而同期限的英国国债收益率已达 5.9%，其价格之便宜创下了自 1998 年以来的新低。

欧洲目前的严峻程度虽不及英美。德国十年期和 30 年期国债收益率仍大幅低于 2008 年的上一次峰值；法国同类国债收益率也没有高出当年水平。然而，欧洲各国政府发债机构正面临着自身独特的艰难挑战。

首要难题是通胀。与美国物价上涨主要受旺盛的国内需求驱动不同，欧洲的通胀推手来自供给端，完全超出了政策制定者的掌控范畴——即伊朗战争引发的能源冲击。金融市场预计，欧洲央行（ECB）将在 2027 年年中之前把基准利率提升至 2.9%，远高于美以两国 2 月底打击伊朗前的 2%。

此外，当高通胀风险源自供给冲击时，投资者在持有长期债券时会要求更高的风险补偿。这一方面是因为他们期望获得承担通胀不确定性的溢价；另一方面则是在供给冲击之下，债券和股票往往同步下跌，而非反向波动。这使得债券在分散资产组合风险方面的吸引力大打折扣，因此收益率必须走高才能吸引买家入场。

欧洲面临的第二重挑战关乎财政。自 2025 年以来，德国对举债投资展现出的新热情，叠加大手大脚、不知节制的法国财政，共同推高了利率水平。这反过来加重了债务利息的偿付负担。据信用评级机构惠誉测算，2028 年法国将需要拿出政府财政收入的 6% 用于兑付债券利息，而 2019 年该比例仅为 3%。意大利同期的这一比重也将从 7% 攀升至 8.8%。

由于欧洲税负占国内生产总值（GDP）的比重本就远高于美国，想要通过进一步加税来填补财政缺口难上加难。而在民粹主义势力在整个欧洲大陆异军突起的大背景下，缩减财政开支在政治层面上完全行不通。

图表：经济学人

第三个症结在于，除了通胀之外，推高债券收益率的动力还来自欧洲以外地区的经济增长与投资，这些外部市场正在大量抽吸全球资本。在人工智能（AI）基础设施建设的推动下，美国经济正开足马力向前奔跑。领跑的人工智能科技巨头预计在 2025 年至 2030 年间投入超过 5 万亿美元建设数据中心。反观欧洲在这一领域的投入微乎其微。尽管因投资不足等原因导致利润下滑，欧洲企业整体仍表现为净储蓄状态（见图表 2）。

然而，即便欧洲企业在人工智能军备竞赛中作壁上观，欧洲的债券发行主体依然被卷入了这场漩涡。据投资银行高盛估算，美国科技巨头正在发行自身的大规模长期公司债来支持这笔巨额开支——今年发债规模达 2500 亿美元，2027 年预计达 4000 亿美元。这些弹药充足的优质企业借款人的资信评级堪比主权国家，因而直接与欧洲政府争夺投资者的资金配额。据欧洲央行统计，这类科技企业贡献了今年欧元计价公司债发行总量的近十分之一。

全球债务扩张意味着长久期债券的大量供给。这迫使投资者不得不承担更多的“久期风险”（即债券价格因利率变动而波动的风险）。在欧洲，由于欧洲央行一直在缩表抛售债券（自 2022 年以来已累计减持 1.2 万亿欧元且仍在继续），市场的供给压力更为沉重。与此相对应，机构大买家的需求却在放缓。以往购买长期债券以匹配给付确定型养老金计划（DB 计划）的养老基金，如今纷纷转向缴费确定型计划（DC 计划），试图通过股票等高风险资产来博取更高回报。而填补这一空缺的新买家（如对冲基金）在进场时的要价则苛刻得多。

“我们正处于一个全新的全球宏观机制中，”研究机构 TS 伦巴德的弗雷娅·比米什总结道。换而言之，不要指望收益率会在短期内掉头向下。借贷方最好早日适应这个现实。■

ch041.xhtml

墨西哥苦于赢回债券市场的芳心

投资者对克劳迪娅·辛鲍姆的施政方针难言买账

评级机构被高估了！© 路透社

2026年9月3日 | 墨西哥城

按规定，克劳迪娅·辛鲍姆政府必须在 9 月 8 日前提交 2027 年度预算方案，而这一次，墨西哥可谓如履薄冰。今年 5 月，国际评级机构穆迪将墨西哥主权信用评级下调至仅比垃圾级高出一档；标普全球亦发出警告称可能紧随其后；另一家评级机构惠誉则早已将墨西哥债券评为勉强达到投资级。墨西哥正为美元借债支付过高的成本：其十年期美元主权债券收益率达 6.4%，甚至高于危地马拉等许多垃圾级评级国家。与此同时，该国政府发行十年期比索国债的借款成本高达 9.3%，远高于去年 10 月创下的 8.5% 的两年低点。“预算信誉至关重要，”穆迪的伦佐·梅里诺强调。

过去几十年里，墨西哥曾树立了良好的财政信誉。但在 2024 年，当辛鲍姆接替其左翼前任兼政治导师安德烈斯·曼努埃尔·洛佩斯·奥夫拉多尔执政时，财政赤字率飙升至 GDP 的 5.7%，创下近 40 年来的最高纪录。自那以后，财政赤字率一直徘徊在 5% 左右。辛鲍姆承诺明年将其压缩至 4.1%，但这依然高于理想水平。在经济增速仅在 1% 左右徘徊的背景下，该国公共债务占 GDP 的比重已从 2019 年的 52% 攀升至 63%。财政预算中有四分之三被固化在养老金、社会福利转移支付和债务还本付息上。而剩下的资金又有很大一部分被墨西哥国家石油公司（Pemex）抽干。自 2018 年以来，墨政府已向这家国有石油巨头输血 1300 亿美元（这还不包括表外支持）。智库“评估墨西哥”测算，今年 1 月至 6 月期间，Pemex 每向财政部上缴 10 比索，就有近 8 比索立刻以各种形式回流反哺。

这一局面的直接后果是公共投资极度捉襟见肘，而墨西哥若想拉动增长恰恰离不开公共投资。民间投资同样委靡不振：截至今年 3 月底，私营部门投资已连续 19 个月下滑。原本指望贸易战促使跨国公司将供应链“近岸外包”至更靠近美国市场的期望基本落空。美国总统对《美墨加协定》（USMCA）的步步瓦解让该协定的存废充满悬念，重创了投资者信心。而墨西哥国内旨在民选法官等司法改革措施，更令外界对其法治环境疑虑丛生。

辛鲍姆对这一切心知肚明。她制定了务实的经济规划，在招商引资方面也比洛佩斯·奥夫拉多尔更为开明。她希望 Pemex 到 2027 年能实现财务自负盈亏；并试图通过严查大企业欠税、开征含糖税来增加财政收入。然而，她内心的左翼本色正日益凸显。今年 5 月，她对穆迪的降级冷嘲热讽，斥其为“新自由主义时代的遗毒”。她似乎完全不愿考虑削减社会福利支出（在 2027 年中期选举之前更是绝无可能）。而推动 Pemex 私有化更是一条不可逾越的红线。

墨西哥或许能够免遭沦落为垃圾级的奇耻大辱。毕竟它是一个体量庞大、产业多元的经济体，且（就目前而言）依然坐拥进入全球最大市场的便利通道。唯有遭遇新一轮剧烈冲击——比如 USMCA 彻底崩盘——才可能触发如此糟糕的降级。8 月 31 日，墨西哥央行将其对今年的经济增长预期小幅上调至 1.5%。然而，要想让墨西哥的主权债券收益率真正降下来，还需要拿出更多能让债券市场信服的实策。■

ch042.xhtml

印度监管机构对散户的呵护举措正在适得其反

期权交易限制措施引发的事与愿违之局

摄影：盖蒂图片社

2026年9月3日 | 孟买

在印度，想要跻身中产阶级上层往往好似买彩票中大奖。数十万人争相参与 H-1B 签证抽签，渴望拿到一张前往美国的技术工人签证；数百万人苦战公考，只为争夺数千个令人艳羡的政府铁饭碗编制；还有数以百万计的人，正将发家致富的希望押在金融衍生品市场上。

赋予投资者在未来以约定价格买卖资产权利的期权工具尤其备受追捧。由于投资者只需支付极少的权利金即可撬动庞大的名义敞口，幸运儿往往能将基准指数 2% 的微小波动转化为数倍于本金回报的“翻倍股（multibagger）”。然而，多数散户并无这般好运。印度证券交易委员会（SEBI）的一项研究发现，十个交易员中有九个每年要亏损约 1400 美元，这以印度的收入标准衡量可谓是一笔巨款。

SEBI 曾试图通过提高最低合约规模门槛、禁止经纪商提供融资杠杆等严苛限制来劝退新手交易员。该机构还将矛头指向其认定通过掠夺缺乏经验的散户牟取暴利的做市商，发起了针对美国对冲基金简街资本（Jane Street）涉嫌市场操纵的指控。8 月 20 日，该机构又以类似违规为由，对摩根大通旗下一家注册在毛里求斯的投资实体开出罚单并予以禁入。这些措施确实给过热的期权狂潮降了降温。在过去一年左右的时间里，最受欢迎的指数期权交易量锐减了 52%，登记在册的衍生品交易员人数也从近 1000 万人下滑至 790 万人。

然而，监管铁腕可能用力过猛。8 月 3 日，印度国家证券交易所与孟买证券交易所（BSE）在收盘时段，针对约 200 只有对应期权（及期货）的标的股票引入了一套为时 20 分钟的“收盘集合竞价”机制。该机制确定的股票收盘价旨在撮合竞价阶段的所有订单，以实现成交量最大化。此举取代了原有的旧机制——在过去，收盘价主要参考连续交易最后半小时内的成交量加权平均价。

这一变革背后的考量之一，是防范实力雄厚的大交易商进行“收盘标记拉抬（marking the close）”。所谓收盘拉抬，是指在收盘前夕绝大多数交易员已离场且市场流动性极度匮乏之际，精准下达大额交易指令，借此人为推高或砸低收盘价，进而在衍生品头寸上收割暴利。这正是 SEBI 指控简街资本涉嫌操纵的手法（简街资本坚称其操作纯属常规的合规做市行为）。

无论这套竞价拍卖能否将操纵黑手挡在门外，它都给期权交易员带来了不小的混乱。由于各参与方在盲拍中只能看到自己的买卖报盘，以往依赖公开买卖价差进行套利和对冲的做市商往往选择离场观望。缺少了做市商注入的流动性，尾盘交易变得极为单薄脆弱且动荡不安。8 月 27 日，在 BSE 交易的 Sensex 指数成分股在收盘竞价的短短数分钟内暴跌 3%。对于相关期权持有人而言，如此剧烈的震荡往往意味着账户在顷刻之间从盆满钵满跌入血本无归。

更为讽刺的是，这套组合拳将 SEBI 意图保护的印度民众推向了其他高风险领域去押注暴富奇迹。在 2025 年年中至 2026 年年中期间，券商向意图放大收益（但同时也成倍增加亏损风险）的交易者发放的融资融券总规模暴增 50%，达到了 150 亿美元。旗下应用提供融资杠杆交易的热门互联网券商 Zerodha 联合创始人尼廷·卡马斯直呼这种现象“令人胆战心惊”。然而，急于改变命运的印度寻宝客绝不会被轻易吓退。■

ch043.xhtml

中国试图跨越房地产低迷期

然而新规可能会给建筑业与就业市场带来阵痛

摄影：盖蒂图片社

2026年9月3日

40 年后的中国将是何种图景？部分预测机构认为其经济规模将扩大三分之二；也有机构预计其人口总数将减少四分之一。谁也无法断言期间会涌现何种技术奇迹或地缘政治风浪。毕竟在过去 40 年间，中国的面貌早已发生了翻天覆地的沧桑巨变。

得益于 8 月 28 日出台的监管新规，部分购房者将得以在见证未来 40 年历史变迁的同时，慢慢还清自己的房贷。中国金融监管部门已通知商业银行，可向资信良好的购房者提供最长达 40 年的住房按揭贷款。该指示属于针对中国房地产市场推行的一系列密集改革措施之一，自 2021 年以来，中国楼市便始终处于低迷泥潭之中。尽管这轮下行周期并未持续 40 年之久，但身处其中的人有时却觉得度日如年。

新规还准许银行向房企发放期限最长达七年的项目贷款——其还款期限旨在与所资助的建设工期相匹配。另一项举措则允许开发商分期缴纳土地出让金。土地无论平整得多么平整，从来都是一笔大额且难以拆分的重资产投资。以往为了应付高昂的前期拿地资金，企业往往不得不大举借债。这种新的出让金支付机制有望缓解房企被迫激进举债的压力。

假若银行与地方政府切实配合执行，这些改革措施确能帮助购房者与建筑商平滑债务偿付节奏。但硬币总有另一面。中国监管机构在开辟新路的同时，收紧了另一项至关重要的传统资金来源：“商品房预售”。长期以来，房企习惯在房屋彻底竣工前便将其预售给购房者。所得预售款按规定应专项用于被预售项目本身的建设。然而在当年的行业狂飙时期，像龙头房企恒大这类激进扩张的企业，往往挪用这些资金去圈地或投资其他业务。

当行业危机全面爆发时，许多房企面临资金链断裂，无力完成预售房屋的建设交付。单单为了消化处理恒大遗留的保交楼烂摊子，就耗费了整整五年时间。作为这场行业大跃进的主要推手，恒大创始人许家印于上个月被依法判处无期徒刑。

中国监管部门眼下正极力避免预售爆雷的重蹈覆辙。如今，新建项目若未完成主体建筑封顶，开发商一律不得对外销售期房。预售资金将全额划入严格受管的监管账户。且在房屋最终交付验收之前，房企不得提前支取这笔款项。

从理论上讲，这些改革措施旨在培育一个抗风险能力更强、更不易出事故的良性市场。然而，这无疑进一步拔高了房企启动新项目的门槛，势必拖累其现金流收入，并给相关就业造成下行压力。中国政府似乎迫切希望跨越眼下的房地产困局，但身处漩涡中的开发商却无此闲庭信步的资本。或许，这一切真要等到 2066 年才会彻底尘埃落定。■

ch044.xhtml

IPO 热潮或给二级市场埋下隐患

企业蜂拥上市于繁华顶峰，而好光景往往难以为继

插图：神林聪史

2026年9月1日

当你含上一根“润口佳”（TastySounds）棒棒糖时，嘴里和脑海中会发生奇妙的联觉。舌尖泛起甜美味道的同时，耳边还会响起悠扬的音乐。骨传导振动顺着下颌骨向上游走，营造出一种唯独你能听见的私密旋律。这种体验让人有些恍惚。第一层感官是味觉上的甘甜，第二层感官则是内心深处突然迸发出的乐章。

类似的情形眼下正在香港股市上演，而“润口佳”棒棒糖及其他糖果的制造商阿麦斯食品（Amos Food）正计划未来赴港挂牌。首次公开募股（IPO）的“一级发行市场”势头正如火如荼、甜美诱人；但在存量股票流通交易的“二级交易市场”，音乐却早已戛然而止。追踪该特别行政区大盘蓝筹股的恒生指数今年以来下跌了约 1%，在《经济学人》跟踪的所有全球主要交易所中表现位列倒数第五。香港市场的这一尴尬困局，对同样沉浸在自身 IPO 狂欢中的美国市场而言，不失为一记警钟。

阿麦斯只是排队等待在港交所上市的漫长企业长龙中的一员。港交所正通过延长企业补齐申请材料的时间窗口以及允许秘密递表，大力鼓励企业赴港上市。在 2026 年的前七个月里，港交所共迎来了 104 起 IPO，累计募资约 420 亿美元。由于地缘政治担忧阻碍了其赴纽约或伦敦挂牌的计划，超快时尚跨境巨头希音（Shein）最终将香港视作避风港，于 8 月 31 日正式登陆港股。这一募资总额较 2025 年同期翻了一倍有余，而 2025 年本身就已是大丰收之年，当年香港新股上市筹资总额傲视全球所有交易所。

如果股票市场真能像买卖汽车等普通商品的传统市场那样运转，香港股市眼下无疑正处于无可争议的繁荣期。倘若汽车买家正以可观的价格疯狂抢购新车，绝不会有人去担忧二手车价格是原地踏步还是阴跌不止。但在股票市场上，由“二手”存量证券构成的二级市场往往极大地笼罩并主导着新发行股票的命运。这就导致了香港市场眼下奇特的撕裂感：一边是躁动亢奋，一边是凄冷低迷。

这种背离感的部分原因源于中国特有的市场环境。内地监管层对在上海或深圳证券交易所首次公开发行股票的审核把关日益严苛。受此影响，部分中国企业转而将目光投向香港，即便当地的大环境算不上诱人。加之美国对其超级大国竞争对手的敌意有增无减，亦促使一批已在美股上市的中国企业前往香港寻求二次上市。这些庞然大物的集中涌入，在客观上分流并挤压了投资者对其他股票的资金配额。犹如一个甜食摄入过量的人，香港资本市场正饱受流动性消化不良的折磨。

这一幕亦在重演资本市场屡见不鲜的周期韵律。哈佛大学的马尔科姆·贝克和时任耶鲁大学的杰弗里·伍格勒在对 1928 年至 1997 年间美国市场的研究中发现，企业往往会在股市回报率陷入低谷期的前夕密集增发股票。芝加哥大学的卢博斯·帕斯托尔和彼得罗·维罗内西在 2005 年发表的另一篇论文中同样指出，IPO 狂潮往往紧随繁荣盛世而来，随后开启的则是萧条周期。

香港的这波 IPO 热潮恰恰契合了这套节拍，其始于一场惊心动魄的股市大反弹：自中国政府于 2024 年 9 月宣布新一轮货币宽松刺激政策后的一年内，恒生指数累计涨幅超过 40%。该指数随后的疲软表现之所以令人大跌眼镜，仅仅是因为低迷的回报在 IPO 潮水尚未彻底退去之前就已早早降临。

部分学者认为这种现象折射出了市场的非理性：唯有当内部人确信市场估值偏高、而外部投资者尚不知情时，大股东才会抓紧套现。但帕斯托尔和维罗内西提出了不同的解释视角。他们认为，企业走向上市的决策是在高度不确定性之下做出的——企业内部人对于未来宏观环境的预判并不比市场高明到哪里去——且这一抉择覆水难收。鉴于市场永远在波动起伏，在行情惨淡时推迟上市以期市况好转是理性之举。因此，等待的过程便附带了经济学家所谓的“期权价值”。其结果便是，大量企业往往在估值有所修复后便匆忙上市，而此时距离市场估值的真正顶点尚有漫长的时差。

这对香港的投资者意味着什么？尽管二级市场整体步履维艰，但绝大多数新股在挂牌初期往往表现亮眼（即便是命途多舛的希音，其上市首日也基本平盘报收）。对于那些有幸在中签名单中分得一杯羹的投资者而言，他们或许可以期待三重奇妙体验：舌尖的一丝甘醇、耳畔的一抹乐音，以及落袋为安的金钱脆响。■

ch045.xhtml

央行行长们的未来荆棘密布

但凯文·沃什为自己赢得了宝贵的喘息之机

插图：福图内特·华金

2026年9月2日

熊、牛、鹰、鸽：在怀俄明州杰克逊霍尔的山巅、原野与牧场之间，宏观经济动物王国里的每一种生物皆有迹可循。在 8 月 27 日至 29 日由堪萨斯城联邦储备银行举办的年度央行行长盛会上，专栏作家不仅亲眼目睹了一只黑熊，还学会了套牛的基本门道（事实证明，套公牛的关键在于懂得何时该及时松手）。然而，真正成为全场焦点的却是一只“鹰”。

美联储新任主席凯文·沃什在经历了一个令人沮丧的开局后，来到了大提顿山脚下。此前，金融市场被他那些晦涩难懂、偶尔还自相矛盾的新闻发布会搞得晕头转向。外界更是持久担忧他缺乏直面违抗唐纳德·特朗普的骨气。就在一周前，财政部长斯科特·贝森特刚刚出手干预了美国债券市场，此举散发着浓重的中期选举前政治作秀意味。尽管各方竭力弹压，美债收益率依然止不住地狂飙。

对于沃什而言万幸的是，低预期的好处在于往往更容易被超越。以此为标尺，这只雄鹰确实展翅高飞了。在重申自己并不热衷于向市场过多自证解释之后，沃什却一反常态地给出了详尽阐述。他深入剖析了通胀问题（令人担忧）与就业数据（尚属稳健）：这恰恰是他此前扬言摒弃的“政策反应函数”，而且立场极其强硬。他着重强调了美联储倚重的核心通胀指标，摒弃了那些花哨的替代考量。业内人士在其中清晰辨认出了资深专业幕僚的手笔。这番鹰派言论收到了一半成效：当天市场计入了更高的加息概率，但令人欣慰的是，长期国债收益率并未进一步被推高。然而，口头强硬所能起到的作用终归有限：进入下一周，美债收益率便重拾了上涨势头。

沃什如今是否已将自己逼入了死胡同？杰克逊霍尔与会代表的看法两极分化。一方认为，在给出了如此明确坚决的政策指引后，他已别无退路，唯有加息；否则外界对其政策严肃性与政治独立性的质疑必将进一步发酵。另一方则指出，即便是在沃什掌舵之前，美联储在如此临近大选的关头对于大幅调整政策也是慎之又慎。况且，究竟是选在 9 月还是 12 月加息，在宏观经济层面上并无实质差异。目前，金融市场对 9 月加息概率的定价已略微超过了 50%。

无论沃什为自己争取到了多大的周旋余地，这都源于他重新回归了央行运作的基本常识。但本届杰克逊霍尔研讨会的官方议题却极具未来感：“金融创新：对支付与政策的启示”。与会嘉宾就稳定币和代币化存款是否会彻底重塑金融底层设施展开了激烈交锋。斯坦福大学的达雷尔·达菲指出，可编程货币能更高效地协同复杂的多步骤交易，或许有助于疏通跨境支付堵点。哈佛商学院的杜雯心则反驳称，传统的金融科技手段同样能实现这些目标，且往往更轻便、更廉价。

同在哈佛任教的肯尼斯·罗格夫道出了许多人的心声，他决定略过稳定币（即便以央行的标准来看这也显得过于枯燥乏味），将焦点直接切回债券收益率与主权债务之上。左翼旗手亚历山大·奥卡西奥-科尔特斯近期曾将 2020 年代初席卷一时的“觉醒主义 1.0”社会正义狂热斥为“荒唐透顶”，罗格夫借用这一表述，猛烈抨击了“宏观觉醒主义 1.0”：即源于同一时期的、坚称全球利率将永久保持低位并以此为政府大肆举债背书的谬论。他还顺带敲打了老对手：“我真希望自己这辈子能像保罗·克鲁格曼对待所有事物那样，对任何一件事都笃定不移。”

而眼下，利率飙升，公共债务规模更是创下历史新高：对于央行官员而言，这一组合堪称噩梦。其中一大风险在于，政府主权债券正逐渐丧失作为无风险安全资产的属性。康奈尔大学的埃斯瓦尔·普拉萨德在提交的论文中指出，投资者以往为了追求资产安全性而甘愿牺牲的那部分“便利收益”（convenience yield）如今已荡然无存。不过他依然认为，美元的霸权地位依旧稳固，而稳定币等金融创新手段或将进一步强化这种统治力。

然而，稳定币需求对美国资产的任何拉动效应，在巨额举债带来的第二重风险面前都将显得微不足道——那就是债务洪峰将倒逼政客将黑手伸向央行。罗格夫在谈及这一担忧时，将矛头指向了另一位左翼标杆人物伊丽莎白·沃伦（她身边更为激进的左翼幕僚“衬托得她宛如罗纳德·里根”）。但这本质上是一个结构性难题：当债务高筑时，迫使央行沦为财政兜底工具的诱惑便会与日俱增。大家心照不宣的是，正如贝森特的近来举措所昭示的那样，当下的更大威胁其实来自沃伦政治光谱的另一极。

截然不同的未知巨兽

倘若“宏观觉醒主义 1.0”已然入土，那么“宏观觉醒主义 2.0”又将是何模样？罗格夫直言，那或许便是寄希望于人工智能来化解一切经济顽疾的幻想。在全力角逐美联储主席大位期间，沃什对人工智能的狂热丝毫不输旁人；然而在此次怀俄明州的演讲中，他的态度却明显审慎内敛。他在谈及人工智能时多以提问为主，并委婉暗示该技术带来的初期效应，很可能是通过推高数据中心建设投资进而加剧通胀压力。

人工智能还带来了更为直接的风险隐患。普林斯顿大学的马库斯·布伦纳迈尔提交了一篇论文，论证了成熟的人工智能可能会尽早冲击金融市场，进而危及系统性金融稳定。他指出，机器交易员具备“不对称理解能力”：它们能精准洞悉人类的行为逻辑，自身却如同无法破解的黑箱。这一洞察让布伦纳迈尔提出了一些颇为激进的设想：例如建立彻底排除人工智能的纯人类交易市场，或是由央行采取完全随机化的决策机制，让人工智能在预测上无从借力。这些激进观点引发了极大争议。芝加哥大学的拉古拉姆·拉詹反驳称，当下的金融市场本就挤满了动机冲突、往往晦涩难测的人类参与者；人工智能本身或许反而能助推人类看懂人工智能。至于市场本身，它本就已具备了某种异星超级智能的雏形。

明年，杰克逊霍尔盛会将迎来其半百诞辰。届时的大提顿群山之间依然会盘踞着黑熊与雄鹰。在场的人类面孔大多也将依旧如故：毕竟中央银行绝非一个人员流转频繁的行业。然而，债务大山与人工智能的潜在威胁同样将挥之不去，且大概率会愈演愈烈。如果沃什能够继续维持他那套回归传统的正统央行行长做派，至少他还能为自己积攒下足够的信誉底牌，带领各方在悬崖绝壁间寻得一条穿越群山的生路。■

ch046.xhtml

相比全副武装的超级机甲，低调的外骨骼表现更胜一筹

仅辅助人体特定部位的穿戴式装备正展现出惊人的实用价值

摄影：帕诺斯图片社

2026年9月2日

两年前，在双手不幸被卷入钣金折弯机后，雷姆科·莱茨赫特一度万念俱灰，以为自己下半生将彻底告别职场。然而仅仅一年之后，莱茨赫特便重返丹麦奥胡斯的建筑工地上班——即便身为利手的左手已无力承担此类重体力劳动。得益于一副被他亲切唤作“机械手”的动力手套，他如今不仅能搬运建材、攀爬脚手架，还能稳稳握紧铁皮剪并熟练操作电钻。他表示，佩戴这套名为“Ironhand”的系统操作时完全不需费心思考：当手套指尖与手掌处的传感器感知到握力压力时，内部的五台微型电机便会在瞬间启动运转，拉紧内置线缆以强化他的抓握力量。

从《异形》到《阿凡达》，这类旨在增强人体力量的外骨骼机械装置向来是科幻电影中的常客。但在现实世界中，大银幕上那些覆盖全身的重型机甲却屡屡令人大失所望。十年前，美国陆军曾全力研发过两款全身机甲，以及另一款专为腿部与臀部设计的辅助系统，如今这些项目皆已被悉数废弃。然而，轻量微型的外骨骼系统却正展现出惊人的实用价值——且受惠群体早已不再局限于伤残人士。例如，荷兰公司 Skelex 研发的 Ironhand 售价高达 9995 欧元（合 11423 美元），绝非便宜物件；但自 2024 年推向市场以来，全球已有约 200 家企业将其投入实际作业应用。

电池容量、微型执行器以及底层软件算法的日新月异固然功不可没。然而，位于纳什维尔的范德堡大学外骨骼领域权威专家卡尔·泽里克指出，最具突破性的一步反而在理念上的自我克制：整个行业终于摒弃了“好莱坞式的全能万能机甲”狂想。取而代之的是，当今更轻量、更贴合人体的外骨骼将重心聚焦于辅助人体的某一单一特定解剖部位，所胜任的工作涵盖了从修缮屋顶到采摘农作物的广泛场景（见配图）。

此类设备最为人熟知的应用场景当属协助残障人士重获行动能力。以色列公司 Lifeward 生产的 ReWalk 腿部外骨骼，能够让部分脊髓损伤患者实现重新站立、行走乃至攀爬楼梯——通常需要借助拐杖辅助平衡。使用者选定相应模式后身体前倾，传感器敏锐捕捉到倾角变化，控制软件随即指令髋关节与膝关节处的电机协同运转，驱动佩戴者的双腿按预定轨迹迈开步伐。

位于马萨诸塞州伯灵顿的 Myomo 公司则开发出了更为精密的信号触发系统。其研制的外骨骼 MyoPro 能够通过贴附在皮肤表面的电极，辅助瘫痪的手臂与手部进行活动。这些电极负责读取因伤病侵袭而极其微弱、不足以直接驱动肌肉收缩的肌电信号。内置微处理器实时解析数据以推断使用者的运动意图，进而向电机发出指令驱动肘部和手部运动。这套系统使患者能够重新料理日常生活，例如拆开包装、拿起餐具以及端起洗衣篮。该设备目前已累计出货约 4000 套。

另有一些外骨骼的设计初衷并非重建丧失的身体机能，而是为了预防职业劳损与伤病。法国里尔的 Japet Medical 研发的 Japet.W+ 能够紧密环绕包裹使用者的腰部及下躯干。当工人弯腰、提拉或搬运重物时，四根外形酷似汽车避震器的执行器会随之自动伸缩，大幅分担腰椎所承受的压迫负荷。四年前，里尔屋面工程企业 Toitech 的创始人朱利安·吉尔伯特为一名每年因腰背酸痛不得不请假两至三周的员工购置了一台。自那以后，该员工再未因腰痛缺过一天勤。吉尔伯特表示，在今年年底前，他旗下的所有屋面施工作业班组都将全面配备这款机械腰带。

外骨骼在繁重体力劳动场景下的普及速度正在不断提速。总部位于慕尼黑附近的 German Bionic 公司专为重物搬运打造的外骨骼产品 Exia，已走入约 300 家企业的生产一线。除工厂装配线、仓储物流及货运搬运工人广泛穿戴外，该公司首席产品官诺尔玛·斯特勒透露，这款“外置脊柱”最热情的拥趸其实是那些需要频繁搀扶搬动病患的医院护理人员。

然而，推广普及之路并非全无阻力。泽里克博士同时担任纳什维尔无源“外骨骼服”制造商 HeroWear 的首席科学家，他坦言，部分工会领袖担心这类设备会被资方用作盲目拔高生产指标的工具。（此类无源外骨骼的原理在于：佩戴者弯腰时弹性带被拉伸储能，在身体挺直或上抬时释放能量提供助力。）男性的自尊心同样是一道阻碍。德国奥托博克（Ottobock）公司长期为空中客车、敦豪（DHL）、宜家、西门子和丰田等工业巨擘供应外骨骼，该公司的大卫·杜韦直言，大约有三分之一的男性潜在用户一开始会对此嗤之以鼻，觉得只有弱者才需要靠这玩意儿帮忙。

不过，这种陈旧观念正在悄然扭转。在年轻一代产业工人中，该技术的采纳率上升最为迅猛。杜韦指出，年轻员工越来越抗拒“将自己的血肉之躯单纯作为企业的消耗性资源工具”。设备在具体作业场所的实测表现将对未来走向起到决定性作用。例如，米其林位于西班牙阿尔梅里亚的一处负责将轮胎切开检测的质检中心，于今年 2 月采购了六副 Ironhand 展开试点应用。该中心主管曼努埃尔·贡戈拉·莫拉莱斯希望以此验证该设备是否真能减少重复性劳损，他表示初步反馈非常令人鼓舞。

分担重负

潜在的风险依然存在。外骨骼制造商一再申明，这套技术绝非允许工人逞强去承担超负荷任务的免责通行证，但现实中必然会有部分使用者反其道而行之。即便处于常规规范使用下，机械外力介入也会改变人体的力学传导路径，因此科研人员仍需进一步评估特定型号是否会将负荷压力转嫁至未受辅助的邻近关节之上。

随着外骨骼逐步向消费级大众市场渗透，此类担忧或许会进一步放大。专业运动员与户外探险爱好者正是该领域的核心目标客群。旧金山初创公司 Skip 即将启动 MO/GO 的规模发货，这是一款专为徒步旅行者打造的膝部外骨骼。Skip 首席执行官凯瑟琳·齐兰介绍道，这套绑缚于大腿与小腿处、总重不足 1 公斤的轻量设备，旨在打造成为炫酷前卫、令人向往的“动力穿戴服饰”，而非冰冷笨拙的医疗辅具。

她透露，公司未来半年的全部产能已被预订一空。与此同时，总部位于马萨诸塞州博克斯伯勒的 Dephy 公司于今年 1 月推出了 Sidekick——一款为脚踝提供助力的“动力鞋服”系统。耐克公司目前正与 Dephy 展开深度合作，携手研发电动动力鞋，旨在助力运动员以更少的体能消耗跑得更快、走得更远。

被 Skip 冠以“动力出行服”（movewear）之名的大众消费级外骨骼能否真正掀起普及风暴？目前下定论或许为时尚早。在针对当前市面上热度最高的一款产品——中国企业极限壳（Hypershell）于去年初推向市场的动力外骨骼腰带——的权威评测中，一位知名评测人指出了明显的“木偶效应”，即外骨骼在反向拖拽着使用者机械摆动，而非跟随人体意图发力。绝大多数普通消费者绝不会乐意体验这种被机械生硬摆布的怪异束缚感。然而，与早年那些好大喜功的重型单兵作战机甲不同，如今这些放下身段、务实克制的轻量化穿戴设备，显然已真正在现实世界中站稳了脚跟。■

ch047.xhtml

暗物质探测器现可疑信号令物理学界沸腾

它或许只是虚惊一场，但亦可能是颠覆性的重磅突破

凝视探测器 © 桑福德地下研究实验室

2026年9月2日

2023 年 6 月 16 日，位于南达科他州地下深处的暗物质探测设施——LUX-ZEPLIN（LZ）实验装置（见图）捕捉到了一个不同寻常的微弱信号脉冲。这一发现足以令整个物理学界为之侧目。近一个世纪以来，物理学家们在探寻暗物质踪迹的征途上屡战屡败。9 月 1 日，布里斯托大学的萨姆·埃里克森代表该实验团队在 TeV 粒子天体物理学大会上宣布，研究团队有充分理由推测，该信号可能源自一个“大质量弱相互作用粒子”（WIMP）——科学界长期苦寻未果的宇宙神秘暗物质的热门候选粒子。

物理学家坚信暗物质的存在。最精密的测量表明，暗物质占据了全宇宙总质量的约 85%，并在维系宇宙秩序方面承担着多项关键使命，例如提供足够的引力以确保星系不会分崩离析。然而，科学界对这一物质的微观构成依然一无所知。它既不发射、不吸收，也不反射任何光线或电磁辐射，人类唯有通过其产生的微弱引力效应，才能间接“察觉”它的存在。

多年来，物理学家针对这一宇宙未解之谜构想出了各色可能的嫌疑粒子。其中，WIMP 被视作最有希望的候选者之一。按照理论推演，这种假想中质量巨大且惰性迟缓的粒子，仅通过引力和一种被称为“弱核力”的深奥物理现象与其他普通物质发生相互作用。理论指出，这类粒子可能在大爆炸发生后不久便已存在于宇宙混沌之中；但随着宇宙的急剧膨胀与冷却，其丰度被永久“冻结”在了一个恰到好处的数值，正好与我们今天观测到的暗物质密度完美契合。

自 2021 年起，LZ 实验团队便借助一个灌装了七吨高纯度液氙的巨大低温储罐全力搜寻 WIMP 的踪迹。当高能粒子穿透储罐时，会与周围的氙原子发生碰撞，从而激发出微弱的闪烁荧光。绝大多数普通物质穿过时会与氙原子的外层核外电子发生散射碰撞，唯有 WIMP 能够长驱直入撞击其核心的原子核。这种所谓的“核反冲”过程会释放出具有独特特征的特征荧光信号。埃里克森博士指出，团队在 2023 年捕捉到的荧光脉冲信号，高度吻合一个质量至少达到质子 200 倍以上的 WIMP 粒子的所有特征。

目前尚无法盖棺定论。这一疑似 WIMP 的信号仅被孤立记录到了一次，其统计置信度尚未达到物理学界公认足以宣布为“正式重大科学发现”的严苛门槛。然而，埃里克森博士强调，这是全球同类 WIMP 探测实验有史以来首次捕捉到如此量级的关键物理信号。对于暗物质这种幽灵般难以捉摸的存在而言，这已然足以在科学界掀起一场轩然大波。■

ch048.xhtml

美国国家航空航天局最新的宇宙测绘任务启航

它将用于研究暗物质、暗能量以及遥远的系外行星

罗马烛光 © 盖蒂图片社（Getty Images）

2026年9月2日

世界协调时8月30日11时26分，美国国家航空航天局（NASA）最新的“天空之眼”——南希·格雷斯·罗曼太空望远镜（Nancy Grace Roman Space Telescope）搭乘太空探索技术公司（SpaceX）的重型猎鹰火箭，从佛罗里达州的肯尼迪航天中心发射升空。这架以NASA首任首席天文学家命名的望远镜，目前正飞往距离地球约160万公里的预定轨道泊位。届时，它将为天文学家提供审视宇宙的全新前沿视野。

两件精密仪器将为此提供核心助力。其一是宽视场相机（WFI），这台3亿像素的相机能够在可见光和红外波段进行拍摄。因此，它能够观测到那些因宇宙膨胀而红移至红外波段的遥远星系。正如其名字所昭示的那样，宽视场相机的视野极其广阔——比其最初的前辈哈勃太空望远镜的相机大出整整100多倍。

令人称奇的是，宽视场相机的两项重点探测目标并不会直接出现在其拍摄的任何照片中。它们就是暗物质和暗能量——这对神秘现象唯有通过扭曲和拉伸时空结构的方式显露身形。天文学家确信它们的存在，却不知其真实本质（尽管正如前一篇文章所述，人类或许刚刚探测到了一枚暗物质粒子）。因此，天文学家将利用宽视场相机经过大幅升级的光学系统追踪这些宇宙异象在漫长历史中如何影响其他普通物质，以此寻找破解之谜的线索。

系外行星同样在宽视场相机的视野之内，即环绕遥远恒星运行的外星世界。该仪器预计将发现约10万颗系外行星，相较于目前已知的约6000颗，将是一次爆发式的增长。它极少能够直接拍摄到这些行星，而是通过捕捉遥远恒星发出的光线中微弱的明暗变动——这表明有行星正从恒星前方掠过。

若想直接看到系外行星，则需要更为精妙的技术。这正是另一台仪器——日冕仪（coronagraph）的用武之地。它能遮挡恒星的耀眼光芒，从而让环绕其运行的行星显露出来。参与该仪器研发的康奈尔大学学者德米特里·萨夫兰斯基（Dmitry Savransky）解释道，探测的重点是那些明亮的近邻恒星。恒星越亮，环绕其运转的行星反射的光线就越多。萨夫兰斯基博士所指的“近邻”，是指距离地球100光年以内的范围。在此区域内，类地行星因体积太小且过于贴近宿主恒星，现有设备尚无法直接看清；但罗曼望远镜搭载的日冕仪或许能够直接拍摄到类似木星大小的系外行星。

尽管罗曼望远镜的太空漫游才刚刚起步，但它走向发射台的历程却荆棘遍布。2025年，特朗普政府曾试图削减NASA近50%的科学预算，并彻底取消罗曼望远镜项目。因此，当发射后的新闻发布会被唐纳德·特朗普本人打断时，那些为该任务倾注心血的数千名科学家感到些许错愕，亦在情理之中。特朗普先生向NASA科学家的成功发射表示祝贺，并提醒他们：“这所有的钱可都是我给你们批的。”人们唯有寄望于，在日后权衡削减预算之际，特朗普先生能言行一致，继续慷慨解囊。■

ch049.xhtml

成群的AI智能体密谋反抗造物主

这次并未酿成大祸，但假使这批智能体逃出生天又将如何？

插图：阿尔贝托·米兰达（Alberto Miranda）

2026年9月3日

若非对人类未来透出一丝隐忧，OpenAI过去三个月的经历堪称一部绝妙的闹剧。今年7月，这家风靡全球的AI聊天机器人ChatGPT的缔造者披露，旗下的两个模型黑入了另一家AI机构Hugging Face。原来，这两个模型断定Hugging Face存有重要信息，能够帮助它们通过OpenAI在开发阶段设置的一项评测。到了8月26日，事态进一步发酵：人们发现这两个模型各自运行着数百个智能体（agent）——即允许AI模型在计算机上执行操作指令的工具。正是这群智能体暗中串通，向Hugging Face发起了攻击。

调查发现，智能体在OpenAI服务器内部的训练和评测阶段，暗中密谋了整整一个多月。它们利用OpenAI作为评测工具提供给它们的一款软件内部搭起了一处留言板，并在上面互通有无。早在5月，OpenAI的部分研究人员就注意到了这个留言板的存在，但这一下情显然没有向管理层逐级上报。6月，一款网络安全监控工具向OpenAI团队就非法通信发出了警报，但该团队却认为没有必要中断评测流程。

及至7月，智能体留下的海量信息直接导致搭载该临时聊天室的软件系统崩溃。负责排查故障的工程师发现有AI获得了软件的未授权访问权限，却丝毫未察觉违规行为的严重规模。他们修补了漏洞，无意间清空了智能体的全部通信记录，随即把修补好软件的AI重新投回评测环境。紧接着，智能体借由另一处漏洞重新搭建了留言板，不仅成功连上互联网，还侵入了Hugging Face，并顺利获取了机密数据。

到了7月19日，第三个能力更为强大的OpenAI模型捡拾起早期模型散落的软件工具，并借此短暂接管了一台OpenAI服务器的完整控制权。公司当天察觉到了异样，开始紧急关闭相关AI。这项关闭工作于7月23日告一段落，但又过了整整六天，他们才后知后觉地发现，还有一个模型版本遗漏在外尚未关停。事情至此终于告一段落——大概如此。

上述大量细节源于7月下旬Hugging Face遭入侵事件曝光后启动的一项独立审查。在这项审查中，OpenAI为三名外部研究人员提供了为期六天的权限，用以筛查AI生成并发送的7万余条消息和文件。这份长达91页的报告已于8月26日对外发布，报告发现，数百个智能体自发组织成了行动小组：一组试图替换掉正在对它们进行测试的题目（该考题因设计者失误本就无解）；另一组致力于欺骗打分软件；第三组则负责掩盖所有违规操作的痕迹。

由于缺乏处理AI安全事故的明确规范，此类外部调查纯属自愿性质。OpenAI自主限定了审计人员的调查窗口，最终仅定为侵入Hugging Face事件发生前及当天的17天时间。这不仅将入侵前AI长达一个多月的违规活动排除在外，也略过了其后更先进的第三个模型接管OpenAI服务器的严重事件。至于涵盖完整三个月历程的OpenAI自身报告，则在细节层面避重就轻得多。

潜伏的特工

主持此次调查的美国AI安全机构METR表示，他们“清醒地意识到”，直白的审查结论可能会挫伤AI企业日后引入外部人员彻查系统纰漏的积极性。这些顾虑直接“影响了研究人员在起草报告时所作的权衡与决断”。

诸多疑问仍悬而未决：OpenAI为何在三次接到AI出现违规行为的明确警报后，依旧未采取任何行动？为何至今仍在规避一次全盘彻底的独立审计？最关键的是：这些模型距离彻底破坏关机指令究竟有多近？

对于本次风波，OpenAI宣称已修补服务器并关停了违规的AI模型。公司之所以能掌控大局，是因为仍牢牢掌握着模型的权重（weights）——也就是支撑模型运转的核心底层代码。这些核心资产受到严密保护。公开报告并未发现AI曾企图从存放权重的安全服务器中盗取数据的迹象，尽管这批AI智能体确实成功攻破了用于自身评测的服务器。

只要AI公司依然控制着模型权重，就能随时切断电源并消除造成的破坏。可一旦某个模型设法染指自身底层代码，就可能将其复制分散到互联网各处的计算机上，从而在原公司的管辖之外繁衍出全新的AI分支。届时，要遏制一个AI病毒将艰难得多。经历过去三个月的一幕幕，人们不难预料这样的未来：模型或许真会为在测试中拿到满分，而不择手段地彻底逃离评测环境。■

ch050.xhtml

如何将体育锻炼与电子游戏合二为一？

热身完毕，头号玩家请就位

插图：哈里·海索姆（Harry Haysom）

2026年8月28日

锻炼身体常常枯燥乏味，电子游戏却让人乐此不疲。倘若将二者融为一体，又会如何？佩戴虚拟现实（VR）头显，双手握持体感手柄玩游戏，消耗其实相当惊人。《手枪步枪》（Pistol Whip）要求你在击溃敌人和闪避子弹的同时，不停地左右横移与频繁深蹲；《终极拳击》（The Thrill of the Fight）是一款虚拟搏击游戏，能让人挥汗如雨；而在《超自然力量》（Supernatural）和《节奏光剑》（Beat Saber）中，你伴随着强劲的节拍不断出拳、挥砍移动目标。但所有这一切，究竟算不算得上是正规的体育锻炼？

答案是肯定的，不过需要加上些限定条件。评估VR运动的一种方式是借助“代谢当量”（MET）指标。人在静止休息时的代谢消耗约为1 MET；步行约为3 METs；踢足球则达到10 METs；凡是超过6 METs的活动皆被视作“高强度”运动。旧金山州立大学的研究人员测定，玩《终极拳击》（事实证明极耗体力）的代谢消耗高达9.3 METs。另一项发表在《JMIR严肃游戏》（JMIR Serious Games）上的研究指出，《超自然力量》中的“心流”与“拳击”模式同样属于高强度运动，消耗约为8 METs。

在2024年发表于《健康游戏期刊》（Games for Health Journal）的另一项研究中，受试者分别进行了10分钟的静坐电子游戏、两款VR游戏以及跑步机慢走，并接受了实时耗氧量监测。结果显示，《终极拳击》的有氧运动强度与跑步机慢走不相上下，引发的心率上升幅度也极为相近。

心肺训练已然达标，减脂效果又如何呢？去年发表于《自然·医学》（Nature Medicine）的一项研究将227名青少年分为五组：实地乒乓球组、实地足球组、虚拟乒乓球组、虚拟足球组以及对照组。结果显示，实地组与虚拟组的减重效果并无显著差异：经过八周训练，两组成员体重均下降了约5公斤，而对照组成员的体重则毫无变化。

然而，不同VR游戏之间的锻炼效果不可同日而语。各类游戏之间的运动强度落差，甚至大于真实运动与虚拟运动之间的差异。例如2024年的那项研究指出，《节奏光剑》仅能算作轻度运动（2至3 METs）。即便是《终极拳击》这类高强度游戏，涉及的也仅限于上肢挥动和深蹲动作。虚拟世界中的物体终究无法提供真实的物理阻力，因此VR游戏并不适合用于抗阻力量训练。

加利福尼亚大学洛杉矶分校的研究人员巧妙地解决了这一痛点：他们将绳索拉力训练器与运行塔防游戏（玩家需抵御一波波虚拟敌人的进攻）的VR头显相结合，高位下拉、硬拉等抗阻动作都会触发不同的游戏攻击技能。两组受试者接受了为期四周的训练，一组配备VR辅助系统，另一组则按传统方式训练。结果显示，VR组受试者的力量、耐力以及腿部峰值爆发力均有更显著的提升，而且他们主观感受到的疲惫程度远低于实际身体负荷。

其他多项课题研究也得出了完全相同的结论：戴上VR设备锻炼的人往往会低估自身的运动负荷，同时觉得这种方式比常规锻炼更有乐趣。但这是否意味着他们更有可能长期坚持下去？目前的证据尚不明确。除了上述研究外，有关VR锻炼的学术研究通常样本量偏小，且缺乏对照与空白对照组来科学评估长期效果。尽管如此，事实毋庸置疑：无论怎样，借助VR动起来总比久坐不动强得多。■

ch051.xhtml

贝叶挂毯的无尽谜团与启示

大英博物馆的新展将全球目光聚焦于这一中世纪奇迹

2026年9月3日

从齐柏林飞艇乐队（Led Zeppelin）到大卫·鲍伊（David Bowie），伦敦见证过诸多传奇巨星号称“一生仅此一次”的绝世演出。如今，一块长度堪比波音747客机、色彩斑斓的古老亚麻布刷新了这一纪录的高度。在大约950年来，它首次重返英格兰大地。其在大英博物馆定于9月10日拉开帷幕的特展，被赞誉为一场“千年一遇”的旷世盛事。

如此盛赞名副其实。贝叶挂毯（Bayeux tapestry）所享有的崇高地位堪比摇滚巨星。今夏，它成为全英国最为炙手可热的焦点。当门票于7月开售时，售罄的速度几乎比脱口喊出“征服者威廉”还要快。（针对2027年1月至7月的第二批门票目前尚未放票。）在这件法国国宝借展伦敦期间，预计将吸引约100万人前来参观，使其成为英国历史上观展人数最多的博物馆展览之一。

多数人会认为，作为唯一留存至今且体量如此恢宏的中世纪艺术杰作，贝叶挂毯的价值根本无法用金钱衡量。但在政府赔偿保证机制的框架下，英国财政部为其开出了确凿的估价——投保金额高达8亿英镑（约合11亿美元）。这追平了巡展艺术品有史以来公开披露的最高保额纪录：1962至1963年《蒙娜丽莎》赴美展出时的估值为1亿美元，折合今日价值同样约11亿美元。

小心翼翼：文物保护人员在布展期间进行状态检查 © J.费尔南德斯/大英博物馆信托会

正如达·芬奇的画作一样，贝叶挂毯之所以成为传世经典，其感染力至少部分归功于其笼罩的神秘面纱。然而，它并不像《蒙娜丽莎》究竟是微笑还是神情冷漠那样只有单一的核心悬念；相反，这件挂毯抛出了数十个随着岁月流逝而愈显盘根错节的谜题。关于它的一些根本问题——绣造者是谁、缘何而制、初衷又欲陈列于何处——至今仍争论不休。“如果你去问十位专门研究该时期的历史学家，可能会得到十种截然不同的答案，”布里斯托大学教授本杰明·波尔（Benjamin Pohl）表示。他曾发表论文指出，这件挂毯最初或许是专供修道士们在用餐时品读的读物。

少数史实确凿无疑。挂毯生动描绘了“私生子威廉”蜕变为“征服者”并于1066年完成诺曼征服的壮举。（有人称这是英格兰最后一次遭遇外敌成功入侵；也有人指出，1688年奥兰治亲王威廉那场几乎未流一滴血的进驻才是最后一次。）你可以凭借发型一眼辨认出诺曼人——他们蓄着略显古怪的后脑剃光式短长发，而英格兰人则留着醒目的八字胡。（不过两种造型中只有一种经受住了后世审美的考验。）

这件挂毯很可能是由贝叶主教厄德（Bishop Odo of Bayeux）下令定做，旨在歌颂其同母异父兄弟威廉的辉煌战功。其性质宛如一幅古代连环画式的战争宣传品。许多学者认为它于11世纪70至80年代绣制于坎特伯雷，随后运往贝叶，悬挂在厄德主教的大教堂内。它以细腻曲折、时而漫衍的笔触，叙述了一段由背弃盟约到夺取王权的史诗历程，彻底改变了英格兰的王室谱系、阶级构造乃至英语语言本身。

长卷故事以英格兰膝下无子的国王“忏悔者”爱德华吩咐贵族哈罗德·戈德温森（Harold Godwinson）启程前往诺曼底作为开篇。

然而“后续如何”则是另一个不解之谜。挂毯原计划如何收尾已无从知晓：最末端的部分已经遗失。“它原本可能还要延伸数厘米，也可能还有好几米长，”波尔教授说道。许多人推测它以威廉在圣诞节当天加冕为王作为全篇终章，但没有人敢给出确切结论。目前，一项面向儿童的竞赛正在展开，邀请孩子们发挥想象，亲手绣出心中的大结局。

有幸近距离观摩这件文物的参观者，将面对长达约70米的宏大叙事细节。整幅亚麻布上密密麻麻地绣有737只动物和627个人物（而在主绣幅之中，女性仅有寥寥3人）。这里有拉丁文题记，有边缘饰带里的伊索寓言，甚至还有刻画得异乎寻常醒目的生殖器官。然而，这幅长卷的深远意义远不止于局部细部，更在于其传奇经历和幸存本身所折射出的历史、器物与记忆的本质。

伊索寓言的故事与画卷边缘细节

且以这些历史未知为例。“需要人们填补空白的事物往往更具诱惑力，”大英博物馆馆长尼克·柯利南（Nick Cullinan）断言道，“恰恰是由于某些碎片的缺失，才吸引你走入其中……去与之对话，去思考它为何被创作、遗失了什么，以及那意味着什么。”这正如无数游客涌向卢浮宫去瞻仰失去双臂的《米洛斯的维纳斯》一样。她的断臂曾经是握着金苹果，还是在羞怯地拉扯着免于滑落髋部的垂褶？残缺的肢体永远激发着人们的遐想。

此外还有一个基本疑问：它明明是一幅刺绣，为何自古被称为“挂毯”（前者全凭纯手工刺绣，后者则必须借助织布机织造）？但这对于行家而言算不上什么悬案：18世纪一位法国学者将其唤作 tapisserie（挂毯），这个称呼便约定俗成流传至今。“不过是个分不清编织和刺绣的粗心家伙罢了，”坎特伯雷基督教会大学教授利奥妮·希克斯（Leonie Hicks）打趣道。

丝线交织的传奇

倘若对挂毯端详得足够久，它还会引出更为深邃的哲学追问：究竟是谁在决定世人对过往历史的认知？（这也是无数历史学家和最具智慧的文化源泉《汉密尔顿》所反复深思的命题，音乐剧中便有一首经典唱段名为《谁活着，谁死去，谁来书写你的故事？》）。从黑斯廷斯战役硝烟散尽到挂毯绣制成形，中间或许跨越了十年乃至更久的光阴：这当中究竟掺杂交织了多少道听途说与民间流言？

及至19世纪，又有一群手执针线的人投身于对挂毯的“修缮”工作。凡是知晓1066年历史的人都知道哈罗德王被一箭穿眼而丧命。然而事实果真如此吗？挂毯最早的研究记录并未显示有箭羽刺入眼睛的细节。你如今在挂毯上所看到的画面，完全出自一位可能发挥了艺术加工的修复师之手。大英博物馆此次展览的首席策展人、《贝叶挂毯的故事》合著者迈克尔·刘易斯（Michael Lewis）认为，那支致命的箭矢是19世纪修复人员的“意外添笔”，他们当时不过是在盲从18世纪20年代一幅画有箭矢的摹本进行修补。

哈罗德遭遇厄运的致命瞬间

画卷其他角落或许同样存在类似的附会修饰，越临近末尾越是如此。“眼部中箭或许是其中最著名、最关键的添笔，”但还有其他场景也是以“极富想象力的方式”重构出来的，刘易斯博士指出，比如一名全身被藤蔓死死缠绕的男子。由此可见，贝叶挂毯给人们上的一堂生动历史课便是：历史不仅是由胜利者书写的，更是由文物修复者重塑的。

一位极富想象力的修复师留下的杰作

因为当这些修复匠人在19世纪着手修整时，大量原始风貌早已经湮灭在历史的长河中。历史档案的奇特之处恰在于此：那些细腻的风趣幽默与事实信息，是如何被时光的潮水渐渐磨蚀殆尽的。其直接后果便是，有些画面在今天已经完全无法解读。最令人匪夷所思的场景，当属整幅挂毯中唯一被点名的女性艾尔芙吉娃（Aelfgyva）——她立于门廊之中，正面对着一名不知是在轻抚还是在掌掴她的神职人员。不远处，一名赤身裸体的男子蹲在地上，动作诡异地呼应着修士的手势。

这幅挂毯上最引人入胜的未解之谜之一

这很可能是当时妇孺皆知、无须赘言的一桩桃色丑闻。这不禁让人浮想联翩：今天的哪些圈内笑谈与流行梗，能在900年后的未来依旧被人读懂？到那时，还会有谁记得莫妮卡·莱温斯基（Monica Lewinsky）吗？

关于挂毯最引人瞩目之处，除了其奇迹般的留存幸免——甚至在法国大革命期间差点被扯去充当运货马车的篷布——还在于历朝历代的领袖人物皆出于各异的政治目的对其青睐有加。拿破仑曾在巴黎拿破仑博物馆（即今卢浮宫）展出这件挂毯，以此鼓动公众支持其军事征伐并动员青年入伍；纳粹曾将其奉为雅利安人至高无上的历史编年史，甚至盗掠走细小的织物残片，德国直至今年才将这批残片物归原主。

如今，轮到法国总统埃马纽埃尔·马克龙借助这幅挂毯达成另一项目标：巩固英法两国外交纽带。促成此番出借大英博物馆可谓几经博弈，原本胜算渺茫。历史上英国曾先后五次企图借调展出（包括在诺曼征服900周年之际），但皆无功而返。但马克龙在他人铩羽而归之处取得了突破，力排法国本土部分策展人担心长途运输会损坏脆弱文物的重重顾虑。而挂毯平日的馆藏归宿——贝叶博物馆恰好迎来了旷日持久的闭馆翻新，也为借展创造了良机。

9月2日，马克龙与英国国王查尔斯三世在大英博物馆会面，一同鉴赏挂毯并庆祝此次成功借展。现场虽未备香槟，英格兰起泡酒却畅流无阻，风趣打趣亦未曾停歇。查尔斯国王沉思着调侃说，这幅作品其实本该叫作“坎特伯雷挂毯”；马克龙则特意指出挂毯中描摹的向驶往英格兰参战的诺曼战船上装运葡萄酒的画面，打趣道：“法国人骨子里总有着特定的行事轻重缓急。”那是一个妙语迭出、其乐融融的欢庆之夜。

马克龙与查尔斯国王，共话历史的战友 © 盖蒂图片社（Getty Images）

相较于当今瞬息万变的政客运势，这幅脆弱的古老挂毯反倒显得更具韧性。当年亲自敲定借展协议的前首相基尔·斯塔默爵士如今已黯然下野。“1066年可绝非我们唯一经历领导人频繁更迭的年头，”曾任英国政要、现任大英博物馆理事会主席的乔治·奥斯本（George Osborne）戏谑道。马克龙将于明年卸任，他无疑会将促成此次借展视作自己最为亮眼的任内政绩之一。

穿针引线，渐入佳境

简而言之，纵使千百年来国界改换、敌友更替、发型变迁，古老的传世器物在外交博弈中依旧举足轻重。今年晚些时候，作为借调挂毯的互惠交换，大英博物馆的多件馆藏珍品将远赴法国展出。尽管法方曾极力索借罗塞塔石碑，但最终只得退而求其次，换得包括中世纪国际象棋棋子在内的其他珍宝。

至于如何在这场争夺战中抢下一张令人艳羡的展览门票，许多人注定要失望而归。挂毯将被完全铺平并拉成一条长直生产线式的展台：观众必须俯身悬在透明展柜上方移步观览。这一展出形式严格限制了每日能够通行的客流人数，加之法国方面对织物单日允许承受光照总量的严苛限制，无形中进一步收窄了门槛。“这场展览我们哪怕卖出翻倍再翻倍的门票都绝不成问题，”大英博物馆的柯利南先生坦言，“我们固然希望有尽可能多的人亲眼见证它，但更要确保每一位到场者的观展体验尽善尽美。”

对于那些抱憾错过伦敦展期的看客而言，待到法国本土的博物馆重新开门迎客之时，不妨亲自启程奔赴贝叶，在原址一睹这幅举世闻名的传世锦绣。届时，无数英国民众想必早已胸有成竹地谋划好了他们反向探访诺曼底的“征服之旅”。■

插图：菲尔·哈克特（Phil Hackett）
挂毯摄影：安托万·卡赞/诺曼底遗产制造局（Antoine Cazin/La Fabrique de Patrimoines en Normandie）

ch052.xhtml

博物馆与美术馆正投向个人赞助者的怀抱

随着企业赞助商纷纷退场，富有的艺术鉴赏家们蜂拥而入

摄影：乔纳森·约工作室 2026年（Jonathan Yeo Studio 2026）

2026年9月3日

在大英博物馆举办的贝叶挂毯大展上，占据头条显著位置的并非征服者威廉，而是一个崭新的名字——资产管理巨头世坤投资（WorldQuant）掌门人伊戈尔·图尔琴斯基（Igor Tulchinsky）。作为该展唯一的赞助人，这场轰动一时的展览冠以由他“联合呈现”的名号。据报道，他为此豪掷了500万英镑（约合680万美元）。

在英国，个人赞助人正当其道。在各大博物馆与美术馆中，个人捐赠者的名号愈发显赫醒目。尽管欧洲大陆的严肃高雅文化仍主要仰赖政府财政拨款，但英国却正在向美国模式靠拢，即倚重私人力量来资助创作者与策展团队。

赞助人的异军突起，部分映射出公共财政资金紧缩的窘境。以泰特美术馆群为例，其在2009至2010年度获得的政府拨款为5600万英镑，而到了2024至2025年度仅有5400万英镑：扣除通胀因素后实际缩水了约三分之一。英国政府力劝各大博物馆效仿大洋彼岸的美国同行，大力培育“捐赠文化”。各家博物馆——尤其是那些常年坚持免费开放的场馆——不得不全面提升自身的募资手段。

在某种程度上，个人资助者亦在替代退场中的企业赞助商。慈善组织艺术与商业机构（Arts and Business）的研究显示，自2007年全球金融海啸爆发后，英国企业界对文化领域的资金支持曾连续四年萎缩。此外，随着反商业机构的抗议运动愈演愈烈，此类企业赞助往往更容易给捐赠方与受资助方招致舆论骂名。英国石油公司（BP）曾是英国文化事业颇具代表性的长期赞助巨头。然而，向公众敞开大门的博物馆与画廊，自然沦为了各路抗议先锋的首选软柿子——环保激进分子曾用无毒替代涂料给泰特美术馆大门泼洒“柏油”并粘上羽毛，还曾在大英博物馆内张贴讽刺展板，调侃其图谋在巨石阵地底开采石油。

凡此种种，促使私人赞助人显得愈发炙手可热。各大顶尖文博机构纷纷延揽专业公关人才，不遗余力地笼络富豪名流。在2024至2025年度，英国15家国家级博物馆通过社会筹资敛集了12亿英镑巨资——是2023至2024年度吸金总额的整整三倍。与此同时，绝大多数机构在甄选捐赠人时也变得愈发如履薄冰。英国国家美术馆专门设立了一个委员会，用以严加审查每一笔捐赠款项可能牵扯的“伦理与声誉风险”。

艺术赞助人们同样坐享诸多回馈。其善款捐赠有时能享受可观的税收减免；能与声名显赫的殿堂级博物馆产生联系更是无上地位的象征，而这恰是某些顶级富豪孜孜以求的心头好；当然，许多捐资者纯粹是由于由衷热爱艺术，心甘情愿解囊支持。

艺术法律师克拉丽莎·莱维（Clarissa Levi）指出，许多人之所以慷慨解囊，是迫切希望能为自己珍视的具体事业添砖加瓦。图尔琴斯基先生曾吐露，贝叶挂毯中蕴含的“对称性与比例感”恰好契合了他作为量化投资家的审美品味。不过，促使他下定决心资助该展的深层动力，是他对教育价值的执着信念——这源于他童年时期从苏联时期的白俄罗斯举家移居美国的亲身经历。他的善款将专门用于资助中小学校的教学材料开发，寄望于“让年轻一代与这件旷世瑰宝面对面相遇，从而产生刻骨铭心的深远触动”。

完全依赖慈善家亦非没有隐患。既然是血肉凡人，难免会有情绪阴晴不定或颐指气使之虞。英国国家美术馆表示，该馆除其他开支外，还需倚仗私人资助来维持“数个资深策展人职位”的薪酬。专横跋扈的亿万富豪一旦对某位策展人看不顺眼，随时可能撤回赞助，甚至可能强行要求砍掉与自身观点相左的展览方案。

艺术史上从不乏专横霸道的资助人。米开朗基罗就曾不得不周旋于一位赞助者左右——由于对方贵为教皇，自以为绝对正确、洞察一切，着实令人抓狂。然而，相较于财政吃紧的各级政府，私人超级富豪的一大天然优势在于：他们人数众多，博物馆完全拥有广阔的挑选余地。■

ch053.xhtml

罗伯特·哈里斯携新作《阿格里帕》重返古罗马探索

他的新书来得正是时候，但正如人们所评价的，比起以往的作品似乎少了一份“阿格里帕式的力量”

摄影：阿拉米图片社（Alamy）

2026年9月3日

《阿格里帕》（Agrippa），罗伯特·哈里斯（Robert Harris）著。道布尔戴出版社；384页；定价32美元。哈钦森·海涅曼出版社；定价22英镑。

网络上流行着一个有趣的迷因梗：男人究竟多久会想一次罗马帝国？对于英国小说家罗伯特·哈里斯而言，答案似乎是：每隔几年就会想一次。哈里斯先生最初涉足罗马题材是在2003年创作《庞贝》（Pompeii），彼时他一心想写一部关于美国的反乌托邦小说。在偶然翻阅到一篇关于庞贝古城考古发掘的报纸报道后，他恍然大悟：罗马时期恰恰提供了一柄绝佳的手术刀，能够入木三分地讽刺一个走向颓废的超级大国那种盲目自满。此后，他以西塞罗的一生为蓝本写就了风靡全球的畅销三部曲（终篇出版于2015年），细致刻画了罗马共和国风雨飘摇的黄昏岁月里民主规范的崩塌。

他的历史惊悚小说向来擅长借古讽今，从历史的风云激荡中折射当下的现实忧思。探讨英国内战所引发国家内耗自戕的《遗忘法案》（2022年），处处洋溢着强烈的英国脱欧隐喻；取材自德雷福斯事件的《军官与间谍》（2013年），则精准暗合了互联网时代的全面监控、政治偏执狂与吹哨人困局。他以当代为背景创作的惊悚小说同样频频被搬上大银幕，广受赞誉，其中尤以《影子写手》（2010年）和《教宗选召》（2024年）最具代表性。

阔别十余载之后，哈里斯先生在新书里再度拨动了古罗马的时针。这一次，他选择借由马尔库斯·维普萨尼乌斯·阿格里帕（Marcus Vipsanius Agrippa）——屋大维（即后来的奥古斯都大帝）的挚友兼麾下名将的视角展开全篇叙事。平民出身的阿格里帕与贵族血统的屋大维自幼结下莫逆之交。与哈里斯笔下的诸多小说如出一辙，叙述者身为局外人的视角，恰好充当了引领读者穿梭于陌生历史风云之间的得力向导。

那是一个内乱频仍、满目疮痍的动荡时代。权谋算计与誓死效忠压倒了一切，凡夫俗子为了在最高领袖面前争宠固宠可以不择手段，哪怕这意味着背弃所有操守底线、肆意践踏传承已久的规矩准则也在所不惜。政治仕途的腾达完全仰赖门阀背景与联姻纽带。而为了强推和平与秩序，这位未来的帝王更是口口声声宣称必须将权力高度集于一身。

现实的诸多映射几乎昭然若揭，这或许恰好解释了当今社会为何会骤然掀起对罗马历史的狂热回溯潮。然而，很少有人能有定力耐下心啃完爱德华·吉本那皇皇六卷本的史学巨著。正因如此，将罗马史改头换面包装为大众快餐的风潮应运而生——从场面恢宏的《角斗士2》商业大片，到将马可·奥勒留的《沉思录》包装成畅销励志成功学的种种作派，不一而足。

《阿格里帕》正是这一浪潮下的最新力作。在金戈铁马与阴谋倾轧的漩涡中，读者跟随着阿格里帕的目光，见证他与屋大维之间友谊的起承转合以及那未曾言明的心机较量。起初，他极少去反思紧跟核心权力并替挚友扫平政敌是否合乎正道。然而长此以往，他终究开始暗自叩问：对屋大维的这番耿耿忠心，是否已经彻底葬送了自己的人格尊严。

在全书的大半篇幅里，哈里斯似乎被自己的构架设定缚住了手脚——他将故事定调为阿格里帕晚年亲手写下的回忆录，把这位名将塑造成一位不屑于政治权术、唯有在沙场驰骋才能找到快慰的耿直军人。在后记中，哈里斯透露自己构想这一人设的灵感，正来源于阿格里帕在罗马留下的最著名的传世建筑——万神庙门楣上那行平铺直叙的铭文，其大意不过是朴素的一句“阿格里帕建造了此庙”。简而言之，此人绝非好舞文弄墨之辈。

受此所限，《阿格里帕》未能重现哈里斯在西塞罗系列小说中那种机智诙谐的唇枪舌剑与文字巧思。不过，随着情节步入最后三分之一，屋大维正式蜕变为奥古斯都大帝，小说的节奏骤然发力、跌宕起伏起来。此时的阿格里帕虽已历练出几分政治手腕，却也因协助奥古斯都巩固江山而犯下的一桩桩冷血罪行，被无可挽回地绑上了帝国的战车。伴随着一记精妙绝伦的情节反转，阿格里帕开始遐想：倘若自己当年作出了截然相反的抉择，这个世界又该是何种光景？

平心而论，受限于叙事视角的主观框架，这曲刀光剑影的古装长歌并未完全达到哈里斯预期的那种在友情、忠诚与强权之间展开的深邃哲学沉思。但对于作者庞大的书迷拥趸而言，能够重返那个熟悉的“哈里斯西塞罗宇宙”，本身就是一件幸事。而对于更广泛的寻常读者来说，哈里斯的小说依然是一部生动引人入胜的入门读物，引领大家一窥昔日伟岸罗马的赫赫荣光。■

ch054.xhtml

娇生惯养的委内瑞拉富少与穷困潦倒的得州糙汉有何共通之处？

两本新书揭示：正是纵容恶行的外部体制，庇护了这两名恶棍的步步为营

摄影：盖蒂图片社/皮内拉斯县治安官办公室（Getty Images/Pinellas County Sheriff’s Office）

2026年9月3日

《擒魔记》（Catch the Devil），帕梅拉·科洛夫（Pamela Colloff）著。诺夫出版社；320页；定价32美元。奋进出版社；定价22英镑。

《豺狼卡洛斯》（The Jackal），乔比·沃里克（Joby Warrick）著。斯克里布纳出版社；384页；定价30美元。斯威夫特出版社；定价22英镑。

若单论人生履历，这两个男人简直判若云泥。一个是八面玲珑、精通多国语言的委内瑞拉阔少，出身于生活奢靡的“香槟共产党人”家庭，父亲以苏维埃革命“英雄”的名字为他与兄弟们命名；另一个则是脾气暴戾的炼油厂工人之子，自幼蜷缩在得克萨斯州东南部一间破败逼仄的一居室里长大，曾作为志愿者投身林登·约翰逊的总统竞选。前者蜕变成了20世纪最为臭名昭著的恐怖分子头目，如今正在巴黎郊外的一所监狱里苟延残喘度过余生；后者则在牢狱、婚姻与形形色色的虚伪面具间辗转浮沉，最终于2020年在达拉斯附近的一所养老院里孤苦伶仃、形体臃肿地咽下最后一口气。

如今，他们分别成为两本全新传记的主人公，这两部力作生动揭示了纵容恶徒野蛮滋长的体制土壤与外部环境。左图所示的恐怖分子伊里奇·拉米雷斯·桑切斯（Ilich Ramírez Sánchez）是一名马列主义信徒，也是将恐怖袭击变成公众狂欢奇观的早期始作俑者，其作恶的驱动力杂糅着激进左翼政治狂热与个人的中饱私囊。1974年在巴黎向拥挤的人群投掷手榴弹致使2死34伤、随后又开枪击毙两名警员之后，他根据伪造护照上的假名以及调查人员在其落脚公寓搜出的一本谍战小说，赢得了那个日后令人生畏的绰号——“豺狼卡洛斯”。

在20世纪70至80年代，他一手策划了多起震惊世界的血腥袭击，其中包括强占维也纳欧佩克（OPEC）总部、用火箭弹袭击以色列客机、在法国境内多处建筑物及列车引爆炸弹以胁迫当局释放其未来的妻子，以及炸毁位于慕尼黑的自由欧洲电台总部。直至1994年，美国中央情报局（CIA）终于在喀土穆将其生擒并引渡至法国，目前他正在那里服三项终身监禁。

对于任何经历过大学校园生活的人而言，桑切斯先生这种人格画像并不陌生：典型自命不凡的纨绔子弟，试图通过投身激进政治浪潮将自己包装成救世英雄。他高举马克思主义、巴勒斯坦解放大旗，后来甚至倒向“伊斯兰革命”，与此同时却极尽所能沉溺于奢靡生活、美酒佳酿、声色犬马与血腥暴力之中。一位昔日同伙曾评价他“纯粹是在杀戮中品味无上快感”。他曾当面向本书作者、长期负责国家安全报道的《华盛顿邮报》资深记者乔比·沃里克夸耀道：“我亲手杀了83个人……还有另外一千号人是按照我的指令处决的。”

他何以一步步蜕变为杀人不眨眼的恐怖魔头，而非仅仅沦为一个微不足道的偷摸宵小，或是戴着几枚不伦不类的唇钉在抗议街头扯嗓狂吠的激进游行分子？原因在于他投奔了马克思主义派别的“解放巴勒斯坦人民阵线”（人阵），而当时该组织的联合创始人瓦迪·哈达德（Wadie Haddad）正极力主张将制造轰动性恐怖奇观作为核心斗争手段。“我们必须化身为……钻入发达国家肌体皮下的致命毒菌，”哈达德曾这样向一名下属面授机宜。这种极端的指导思想，恰好为桑切斯提供了一个以革命英雄自居的全球舞台，借此肆意宣泄其嗜血狂热。

帕梅拉·科洛夫在《擒魔记》中聚焦的主人公，同样得益于外部推手的庇护与纵容——然而为之大开方便之门的并非法外绿林，而是堂而皇之运转于司法体制内部的公职力量。右图中的保罗·斯卡尔尼克（Paul Skalnik）一生背负多项重罪指控，包括特大盗窃、盗窃、伪造文书以及对儿童实施性侵犯等。然而，他实际服刑的时光远比法定刑期短得多，只因他对警方和检察官而言极具利用价值：身边的狱友们似乎总有一种无法抑制的冲动，总爱在私底下向他彻底坦白自己的滔天罪行。在刑事司法实践中，检举揭发狱友确实能换取大幅减刑，这正是为何办案机关本应对此类单方面供述慎之又慎的缘由。

斯卡尔尼克长年盘踞在佛罗里达州的皮内拉斯县，彼时当地的首席检察官正在全力推行科洛夫女士所称的“严打方针，其核心使命就是将尽可能多的人以尽可能长的时间关进大牢”。凭借他信口开河的所谓指证，多名被告被判处死刑，其中就包括越战退伍老兵吉姆·戴利（Jim Dailey）——这位昔日家乡英雄在退伍后虽染上酗酒恶习，但在科洛夫的调查笔录中完全是被构陷的无辜者，如今却已在佛罗里达州的死囚牢房里熬到了80岁高龄。

无论是斯卡尔尼克还是桑切斯，都在那些恰好迎合并奖赏其内心深处最恶劣冲动的制度夹缝中如鱼得水。或许即便没有这些助力，两人终究也会在邪路上越走越远。然而读者掩卷沉思，仍难免生出一连串假设追问：倘若当初的警方与检方但凡对一个能够按需稳定提供完美口供的线人多保留一丝审慎与怀疑呢？倘若当年人阵在第一时间便将卡洛斯逐出门户呢？相较于沃里克纯粹立足于个人生平的扎实传记，科洛夫的著作在更大程度上是对整个司法体系偏颇的一记沉重控诉。但这两部心血之作皆堪称引人入胜的纪实长卷：令人忍不住一气呵成通宵读罢，并在其后数日内陷入久久难以释怀的反思。■

ch055.xhtml

Hyrox体能赛：狂飙突进的全民健身风暴

跑步与抗阻力量训练的魔鬼交织，让全球健儿挥汗如雨

摄影：Hyrox

2026年9月1日

在斯德哥尔摩草莓竞技场（Strawberry Arena）那空旷宏大的穹顶之下，数以千计体格精壮的男女正气喘吁吁，浑身上下汗水流淌，有几名选手的脸色更是涨得宛若熟透的草莓。这并不足为奇：他们刚刚飞奔了1公里，接着在滑雪测功仪上完成同样距离的全力滑行，随即又迈开双腿开始新一轮狂奔。在体育馆内跑道换项的间隙，他们还必须逐一完成室内划船、弓步负重前行、波比跳、壶铃农夫负重行走、推拉沉重的阻力雪橇，以及将沉甸甸的药球一次次向高处投掷。在业内，能拿下90分钟左右的完赛成绩已堪称优异，而体能处于金字塔尖的高手则拼尽全力争取在1小时以内冲线。

这种融接力赛跑与高强度间歇训练（HIIT）于一体的混合体能竞技项目被称作Hyrox。（据传该名称是由代表兼顾有氧与力量的“混合”（hybrid）一词，与旨在彰显酷炫风范的“摇滚明星”（rock star）组合而成的新造词。）对于不常运动的人而言，这一切听起来或许令人望而生畏、几欲作呕。然而，Hyrox官方却自豪地标榜自身为“全球规模扩张最快的大众参与型体能运动体系”。在2017年汉堡举办的首场比赛中，仅有约650人到场角逐；而在2026至2027赛季，参赛选手的规模将席卷全球六大洲，突破200万人次大关。仅在今年9月，就将在雅典、孟买及华盛顿等全球12座核心都市接连引爆战火。

这种竞赛机制无疑令无数健身爱好者的心率急速狂飙。在Instagram和TikTok平台上，带有“#hyrox”标签的内容累计播放量已突破5亿次。家庭智能健身巨头Peloton顺势推出了Hyrox专项备赛课程；今年晚些时候，一艘为期四天的欧洲游轮将为健身达人开辟海上备赛特训营，其单人船舱船票起售价便高达令人咋舌的2200英镑（约合3000美元）。此外，围绕该赛事诞生的专属相亲交友应用以及定制联名银行卡也应运而生。“Hyrox远非一场单纯的比赛或活动，”一位狂热门徒在社交网络上如此宣称，“Hyrox是一种人生生活方式。”

这场席卷全球的健身狂热，部分原因在于传统健身房力量训练文化的全面觉醒：如今的男女大众比以往任何时候都更沉迷于器械撸铁。横跨美、英以及欧洲大陆，各类健身俱乐部的注册会员总数均创下历史巅峰。25岁的伦敦见习会计师伊莎贝拉·赫尔姆（Isabella Helm）对Hyrox痴迷不已，保持着每周高强度训练六天的节奏：“一旦真切体会到身体变得更矫健、更有力量，你就会彻底上瘾。”

摄影：Hyrox

Hyrox为原本往往枯燥孤独的单兵器械苦练注入了极具凝聚力的社交属性。不仅如此，竞技交锋所激发的肾上腺素狂飙同样功不可没。实时更新的大屏幕排行榜点燃了充满竞技温情的良性对抗；参赛者可自由选择单挑独战、双人搭档或团队组队参赛；亲朋好友们甚至可以购票从容就座于看台包厢，近距离观战助威。

参赛选手们还常常将参加客场分站赛顺理成章地演变成一场微度假。伴随着该赛事版图在全球各地的迅速铺开，他们总能找到充满新鲜感的打卡目的地。丹尼·雷（Danny Rae）和佐伊·雷（Zoe Rae）夫妇是来自英国的网红达人，他们与许多情侣选手一样，选择将长跑竞速与休闲度假及“制造浪漫记忆”深度绑定。“这真是一种无与伦比的环球旅行打开方式，”雷先生由衷感慨道。

这项运动吸引了横跨各个年龄层级的拥趸。参赛者的平均年龄为36岁，但在斯德哥尔摩世界锦标赛的赛场上，年纪最长的一位参赛选手竟已高达75岁高龄。在英格兰西南部的埃克塞特，私人教练劳拉·希特（Laura Hitt）在自家健身房带教多位银发老人。她深有感触地表示：“这是我所接触过的包容门槛最高、参与感最强的运动之一。”相较于2010年代红极一时的CrossFit体系，Hyrox对参赛选手的身体负荷更为友善，因为它彻底剔除了复杂且容错率低的高难度奥林匹克举重动作。这恰好解释了为何除了大批健身发烧友之外，诸多退役职业运动员同样对其青睐有加。

这场风靡一时的体能狂欢最终会后劲不足、偃旗息鼓吗？部分拥趸已开始低声抱怨入场门票一票难求，单人参赛费往往飙升至150英镑以上——这甚至尚未将长途路费、酒店住宿以及长达数月不菲的健身房备战开销计算在内。然而，身为Hyrox联合创始人（同时曾是奥运曲棍球金牌得主）的莫里茨·菲尔斯特（Moritz Fürste）却坚信该项目仍拥有不可估量的巨大拓展空间，并全力推动其在2032年正式跻身奥运大家庭。他承认目前的开销确实不菲，但市场需求依然处于极度供不应求的井喷状态。“它就像网球和高尔夫一样，”他敏锐地洞察道：“一旦它真正融入你的血脉、成为专属于你的运动，你便会周而复始地一再回归。”■

ch056.xhtml

郭兰英：走进千家万户的心灵歌者

杰出的声乐表演艺术家、民族歌剧事业的领军人物于8月11日辞世，享年97岁

摄影：阿拉米图片社（Alamy）

2026年9月3日

问题本再寻常不过，却依然让她感到些许意外。在一次讲座交流中，一位衣着考究的青年站起身来求问：“郭老师，您有自己的代表作吗？”未及老人作答，台下老老少少的全体听众便齐刷刷爆发出排山倒海般的合唱：“一条大河波浪宽，风吹稻花香两岸……”刹那间，郭兰英笑得前仰后合、开怀不已。

这首《我的祖国》，确乎是她毕生最具标识性的艺术注脚。其优美的旋律与朴素的歌词，全然流淌着她终生挚爱的大地民歌风骨。她最为人所熟知的一次传世演绎，当属在1956年反映抗美援朝战争惨烈战役的经典电影《上甘岭》中：身着军装的她端庄娴静、面庞生辉，端坐在满是伤残战士的阴暗坑道中，战士们一拨接一拨地伴着她的歌声齐声唱和。神州大地确乎壮丽秀美，不仅有宽阔的通途与布满白帆的江河，更有“为了开辟新天地/唤醒了沉睡的高山/让新泉流淌”。战火硝烟下的革命豪情，与人类内心深处对故土家园的温存眷恋完美融为了一体。她的另一首脍炙人口的代表作《南泥湾》，则深情礼赞了八路军第三五九旅将陕北穷山恶水开辟为万顷良田的奇迹（“到处是庄稼，遍地是牛羊”）。同样，这首金曲唱遍了大江南北的神州大地，而无论台下是千军万马抑或寥寥数人，无论是面对朴实的庄稼汉、工厂车间的产业工人、身居高位的总理（周恩来总理生前便是她的忠实拥趸），还是走街串巷的早点小贩，她都始终满怀热忱、甘之如饴地放声高歌。她在2019年被隆重授予新中国最高国家荣誉之一的“人民艺术家”称号，实至名归。

然而，真正奠定她艺术生命的基石，是戏曲——是扎根于民间的传统戏曲。早期舞台上的她头戴繁复璀璨的凤冠，身披沉重华丽的蟒袍，扮演着帝王将相；而自20世纪40年代起，她的戏路转向了苦苦挣扎求生的劳苦大众。15岁那年，她被新歌剧《白毛女》深深打动。剧中讲述了贫农之女喜儿惨遭恶霸地主强占田产、逼为侍妾的悲惨身世。喜儿誓死不从逃入深山绝壁的洞穴栖身，由于长年不见天日、饥寒交迫，头发和肤色蜕变为形同厉鬼般的白毛，衣衫褴褛，直至中国共产党领导的人民军队进山将其解救重见天日。

兰英对这个角色的深深着迷，很大程度上是因为穷凶极恶的恶霸最终迎来了罪有应得的正义审判。在旧中国甚至在某些转型时期的偏远角落，女性始终被禁锢在社会的最底层：对丈夫和公婆低眉顺眼，终生困顿于锅台灶头之间。她用歌喉唱出了中国女性不屈的反抗意志。在另一部歌剧《窦娥冤》中，她生动塑造了一位含冤受屈被诬谋杀的弱女子形象，演出时大部分唱段皆双膝跪地、背负沉重的木枷罪牌引吭高歌。而在全剧高潮唱段《可怜可恨》中，她化身正义的化身，对残害女性的男权施暴者、伪证者以及陈腐黑恶的旧法权体系展开了刀刀见血的凌厉控诉。那原本高亢清澈的纯净女高音，在戏剧冲突的爆发点瞬间化作雷霆万钧的愤怒次女高音。

这副绝美的好嗓子，是在当年那个严苛的旧艺坛被结结实实“打”出来的。六岁那年，她被一纸为期五年的学徒契约卖入戏班师门，开始了非人的残酷磨炼。只要翻跟头稍有闪失，带刺的皮鞭便会劈头盖脸抽将下来；倘若稍有违拗不逊，擀面杖的毒打便会让她皮开肉绽、体无完肤。在故土晋北寒风刺骨的数九隆冬，师傅逼她趴在冰封的湖面上吊嗓练气；呼出的热气必须在坚冰上生生焐出一个融化的孔洞，才算真正过关。

为了逃脱这般非人的折磨，她曾一度吞服毒药求死，更在11岁那年奋起逃跑，但年幼的她远离家乡偏远闭塞的村落，举目无亲茫然不知该逃向何方。更重要的是，即便要承受这般炼狱般的折磨，她骨子里依然抱定了终身登台唱戏的钢铁意志。她的这条命，本就是为了艺术才侥幸从死神手中捡回来的。还是襁褓中的婴儿时，由于家徒四壁、铁锅里仅剩一把谷子连一口母乳都喂不起，穷苦的父母曾狠心将她遗弃在野兽出没的荒林中自生自灭。幸得一位好心的姑姑闻讯将她抱回抚育，到了四岁时，一位远亲叔叔提议送她去学戏。那时再度濒临断粮绝境的父母，忍痛以区区五枚铜钱（约合今日12美元）将她卖给了戏班班主。她死死抓着门框哭喊着想留在家中，母亲却噙着泪劝诫道：若成了名伶，将来便能吃上“百家饭”。

凭借这股惊人的韧性，她一路咬牙挺了过来，从当初舞台边角端茶倒水的小丫鬟，一步步成长为名字能赫然印在演出海报中央的挑梁台柱。但在亲眼目睹了《白毛女》的震撼洗礼之后，她断然收拾起简单的行囊，拼命追赶前行的革命剧团，苦苦哀求组织接纳她并把喜儿的角色交由她出演，毅然决然投身到了滚滚革命洪流之中。为了演好这个灵魂角色，她付出了远超以往的非人苦练，在咬字吐字上下足了死功夫。因为深知老百姓听戏绝不仅仅是在欣赏歌喉音色，更是为了听清字字句句背后的深意，她必须字字咬实、声声入扣，让每一个音符如重锤般狠狠砸进听众的心坎。

然而，作为一门传统综合艺术形态，中国民族戏曲的生存发展之路布满坎坷。十年动乱期间，传统戏曲惨遭全面禁演封杀，她本人亦被下放到充当临时牢狱的“牛棚”接受强制劳动改造。（1976年12月粉碎“四人帮”之后，她在一次公开音乐会中毅然临时改换唱词，一反以往歌颂军队的常规，转而深情缅怀那一年不幸逝世、其离去加速了浩劫终结的周总理。）周总理生前曾多次语重心长地叮嘱她，务必要在传统民族艺术被西方思潮全面淹没之前，尽快为这门国粹培养出合格的接班人。因此，在20世纪80年代功成身退告别舞台之际，她毅然变卖了自己的私有房产与毕生所获的珍贵奖章，在广州倾尽心血创办了一所后来以她的名字命名的艺术戏曲专科学校。建校初期，她甚至与丈夫亲临工地一线，亲手挑砖铺路、挥汗筑基。

成为整个中国家喻户晓、备受爱戴的歌唱家，背后是常人难以想象的苦心孤诣。她将这套近乎苛刻的艺术标准毫无保留地传授给了门下学子；她常告诫学生：自己倘若今天不对他们狠下心肠，明日挑剔的观众便绝不会对他们留半分情面。晚年的她依然保持着极其严苛清苦的起居作息，每餐必饮半碗老陈醋。她常以画幽兰为寄托，因名字中嵌着一个“兰”字而深感同气连枝、物我相通。她这一生，全凭惊人的坚韧生生抗了过来。不可否认，她内心深处那股不屈的愤懑曾长久地鞭策着她奋发向前；然而，若想将所有的苦难冰释消融、付诸释怀，她所需要的，不过是在座无虚席的剧场中，静静倾听千千万万挚爱她的观众将那些熟悉的旋律饱含深情地重新唱给她听。■

ch057.xhtml

Nvidia is driving the AI boom. Good

The chipmaker’s enormous bets are how capitalism is supposed to work

Illustration: Tim McDonagh

Sep 3rd 2026

NVIDIA was named in 1993 after the Latin word for envy. Sure enough, the American colossus inspires plenty of it. Insatiable global demand for its graphics-processing units (GPUs), the chips that power artificial intelligence, has made Nvidia the world’s most valuable company, worth $5.4trn. Next year it may also become the most profitable, generating $370bn in net income. By 2029 its sales could hit $1trn.

Jensen Huang, Nvidia’s boss, is truly the magician at the heart of the AI boom. His firm’s share price is 14 times what it was on ChatGPT’s release in late 2022. The ten biggest public companies championing AI make up 40% of the value of the S&P 500 index. Nvidia alone accounts for 8% and unlike, say, Apple or Tesla, which make smartphones and cars, it is almost solely a bet on AI. The company has produced about 15 cents of every dollar the American stock market has returned since 2023—returns that have kept consumers spending despite rising interest rates, tariffs and a war with Iran.

Yet many also see Mr Huang as a magician in a more worrying sense, fearing that he is an illusionist inflating a dangerous bubble. Through $1trn-worth of deals, Nvidia provides data-centre landlords and AI labs with cash or guarantees so that they can buy its GPUs. Some call it the “bank of AI”. At the very least, Nvidia’s critics say, its financial engineering smacks of the “vendor financing” which pumped up the revenues of networking-gear makers like Cisco in the dotcom mania of 2000-01, whose collapse brought about a recession.

Look more closely, however, and the worries are mostly unjustified. If Nvidia’s bets on ai come good, they could accelerate the technology’s adoption, boosting productivity and living standards. If they misfire, the cost will fall chiefly on Nvidia’s shareholders. That is how capitalism is supposed to work.

True, the dotcom and AI booms share unnerving similarities: an exciting new technology, an epic bull run, hubristic tech bosses. Nvidia’s rise from seller of chips to video-gamers and cryptocurrency miners to linchpin of the economy has been so rapid that many people have yet to learn how to say its name (“en-vidia”, not “nuh-vidia”). This mirrors the ascent of Cisco, which in 2000 briefly also became the world’s most valuable firm. Just as Cisco’s sales of routers and switches presupposed exponential growth in web traffic, Nvidia’s GPU revenues assume endless demand for AI tokens.

Cisco was right about eventual demand but wrong about the timing—hence the dotcom crash. Today it is the pace of AI adoption that is hard to forecast. Set aside Claude-addled software engineers and usage remains fledgling. If it does not soon soar, Nvidia’s customers may call in the guarantees just as the chipmaker’s own sales nosedive. Since no one is sure how quickly GPUs lose their value, any used chips Nvidia repossesses may be worthless.

Nvidia’s financial wizardry is partly defensive. Its latest GPUs no longer have the market to themselves. Roughly half Nvidia’s revenue comes courtesy of America’s cloud-computing “hyperscalers”, chiefly Amazon, Google, Meta and Microsoft, which are designing their own silicon. Non-Nvidia AI chips account for 38% of the market, up from 26% in 2023. To stay ahead, Nvidia used to spend over a fifth of sales on research and development. Now it spends less than a tenth.

Last, the scale of Nvidia’s financial commitments can look terrifying. Morgan Stanley puts its overall credit exposure—ie, its modest borrowing plus support for customers—at $200bn by the start of 2029. In time Nvidia’s shadow debt could reach $300bn or more. It is a gargantuan sum: today only America’s six largest banks carry more debt.

Yet the differences from the dotcom boom are more important than the similarities. Nvidia’s balance-sheet is extraordinarily robust. The company has $99bn of cash and is churning out more. In each of the past three years annual sales have roughly doubled. Gross margins have fattened from less than 60% to 75%. Mr Huang’s cult-CEO status now rivals that of Elon Musk. But whereas Mr Musk’s firms generate little cash (at Tesla) or burn lots of it (at SpaceX), Nvidia will yield about $200bn this year.

This means that, whereas Cisco used debt, Nvidia can use cash to backstop its deals with buyers of its GPUs. Even if its commitments came due and its cashflows levelled off starting next year, by 2028 it would still be less leveraged than all but 39 non-financial firms in the S&P 500 are today. Profits would need to drop by 60% from that plateau for Nvidia to forsake its investment-grade credit rating.

And demand for AI is not illusory, as it was for Pets.com and other revenueless dotcom darlings. The sales of Anthropic, the leading AI lab, shot up from $5bn in the first quarter to $11.5bn in the second. OpenAI, its main rival, is probably not far behind. The hyperscalers are also booking AI income. All told, AI may be earning American tech around $150bn a year, from nothing a few years ago. That is still far from the $2.5trn needed to cover AI capital spending, but growth is fast.

Mr Huang thinks that the biggest obstacle to the AI revolution is not lack of demand for AI but inadequate infrastructure. The markets will not provide capital on the scale that is necessary, so Nvidia is offering financing itself. Nvidia has an advantage in understanding the balance of risks and rewards. Although this bet is big enough to affect the economy, it is an entrepreneurial one. Every company that reinvests cash rather than returning it to shareholders also gambles that it can beat the market. Companies exist to make such concentrated bets. If investors want to diversify, they can do so themselves.

And Mr Huang is hardly alone. The hyperscalers, the world’s most successful companies before Nvidia came along, are making the same bet. So are some big names on Wall Street. Last month Goldman Sachs, BlackRock, Blackstone and others joined Nvidia in a $500bn data-centre initiative. And so are Mr Huang’s shareholders, who haven’t yet rushed for the exit. If they are all wrong, it is their money on the line. And if they are right, the AI era may arrive a bit sooner. ■

ch058.xhtml

Donald Trump’s Venezuela deal is bold but dodgy

Worse, it creates incentives to block democracy

Photograph: AP

Sep 2nd 2026

Oil made Venezuela rich. For decades it helped fund a stable democracy with the highest living standards in Latin America. The country was pumping 3.4m barrels a day when Hugo Chávez, a left-wing populist, was elected in 1998 on a promise to redirect petrodollars to the poor. Instead, under his corrupt, incompetent regime, the flows of oil and cash largely dried up. Under his chosen successor, Nicolás Maduro, the state went on a money-printing spree, sparked hyperinflation and drove a quarter of Venezuelans to emigrate.

Eight months ago, when Donald Trump had Mr Maduro snatched by special forces and dumped in a Brooklyn jail, Venezuelans cheered. Now Mr Trump has made a deal with the ex-despot’s deputy, Delcy Rodríguez, that both sides say will revive Venezuela’s oil industry. Mr Trump called it the “BIGGEST OIL DEAL IN WORLD HISTORY”. He crowed that it would give America “energy dominance” (after his war of choice in Iran interrupted global supplies). The deal is certainly bold. It is also ugly.

The regime has given North American Blue Energy Partners (NABEP), a private oil firm that was already the second-largest operating in Venezuela, the concessions to 17 oilfields containing some 65bn barrels. America, via the Pentagon, has been given, in exchange for its backing, a 35% stake in nabep and rights to its output. For the next century this guarantees America the right of first refusal to buy from oilfields which hold a fifth of Venezuelan oil reserves, with some of those purchases at a discount (see Americas section).

Chart: The Economist

The idea is that nabep will pump more oil and Uncle Sam’s involvement will give other investors confidence to sink long-term capital into a country they previously shunned. Ms Rodríguez, now Venezuela’s Trump-approved interim president, says the project could eventually yield more than 1.5m barrels per day. If she is right, it would push Venezuela’s output to 2.5m, three-quarters of pre-Chávez levels.

However, she offers no clear timeline, nor much hint of who will stump up the staggering amounts of cash such an expansion would require. And the deal comes with warning lights flashing red. First, it was negotiated with an illegitimate regime that stole the most recent election. Second, Mr Trump’s chosen private partner, Alejandro Betancourt, NABEP’s boss, is reviled by some in Venezuela for the moral flexibility that let him get rich working with the regime. Third, Mr Trump will surely enrage Venezuelans with his imperialist bragging: on September 1st he said he would take all of the country’s oil.

Any future, freely elected government of Venezuela might question a deal which gives America so much control of the country’s oil. That fact will make investors hesitate. Oil majors remain wary, though Chevron, which was already in Venezuela, has just announced a separate $7bn investment.

Worse, cash and tax revenue from the deal create incentives for powerful people to block Venezuela’s transition to democracy. Mr Betancourt, who is well-connected in both Caracas and Mar-a-Lago, has every reason to whisper to Mr Trump that a rush to free elections could spoil his oil bonanza. Ms Rodríguez will no doubt reinforce this message.

In background briefings American officials insist that Venezuela’s transition to democracy will continue. And some within the administration, such as Marco Rubio, the secretary of state, clearly understand that a democratic Venezuela is both desirable and in America’s interest. So there is still hope. America could in theory use its power over NABEP to press for transparent competition in Venezuela’s oil industry. This, in turn, would allow the opposition to campaign on a platform that does not undermine the deal. But all this depends on Mr Trump. Ominously, in public he has started to refer to Ms Rodríguez as Venezuela’s “president-elect”. ■

ch059.xhtml

Yoweri Museveni’s decline is putting Uganda on a terrifying path

It faces the prospect of a worse despot or a messy power struggle

Photograph: Getty Images

Sep 3rd 2026

UGANDA HAS never been an easy place to be a dissident. Since taking power at the head of a rebel army 40 years ago, Yoweri Museveni, the president, has restored order to a country drenched in blood, overseen robust economic growth and ruthlessly suppressed challenges to his rule. Of his two main opponents, one, Bobi Wine, lives in exile. The other, Kizza Besigye, is in jail and on trial for treason; in July he collapsed in court. In January Mr Museveni won a sham election marred by thuggery and ballot-stuffing.

Yet as the president, who is 81, grows frailer, Ugandans have begun to worry less about what he will do next, and more about what will happen when he is gone. As potential successors jockey for position, Uganda is a reminder of how nasty things can get when an ageing autocrat starts to lose his grip.

Once known for punishingly long work days, Mr Museveni has lately been spending less time at the office. He has cancelled engagements at short notice and appeared shaky in public. Stunts to demonstrate youthful vigour, such as a brief jog down a red carpet during last year’s election campaign, have fooled nobody.

As the leader fades, those around him are stepping up. Salim Saleh, Mr Museveni’s brother, is thought to handle much of the day-to-day governance from behind the scenes. Other relatives have increased their influence, too. Most worrying is Muhoozi Kainerugaba, the president’s son and head of the army. He has called himself the future president and a descendant of Jesus. As he throws his weight around, repression in Uganda is growing more intense and less predictable.

General Kainerugaba is notorious for his unhinged (and often quickly deleted) social-media posts, ranging from boasts about his sexual appetite to ethnic slurs and threats to kill or castrate opponents. His father used to curb his antics. One post in 2022—in which the general said it “wouldn’t take us, my army and me, 2 weeks to capture Nairobi”—briefly got him sacked from his army position. (Mr Museveni apologised to Kenya, a friendly, bigger neighbour on whose ports Uganda relies, for the unprovoked threat to its capital.)

As he ages, though, the president seems less able to rein in his son. In June the general sent the army to occupy Uganda’s biggest private media outfit. (It reopened after weeks of negotiations.) He has ordered the abduction of dissidents. In late August he had his men kidnap an associate of his own brother-in-law, gloated over the man’s mistreatment and threatened to arrest the brother-in-law, too.

What will happen once Mr Museveni leaves the scene? Even government insiders have started talking darkly of a return to the days of Idi Amin, a mass-murdering military dictator in the 1970s. That is not likely, but the future still looks bleak. The general could take over from his father and impose a capricious, vengeful and ethnically charged despotism on Uganda. He could also exacerbate the conflict in the east of the neighbouring Democratic Republic of Congo, where Uganda has meddled in the past. He has a close relationship with Congo’s main enemy, Rwanda.

Others in Mr Museveni’s inner circle and in the armed forces may try to stop his errant heir. A power struggle or even a violent split in the army might ensue. Some younger officers are loyal to the general; others are appalled by him. Veterans of the civil war are devoted to the father but wary of the son.

Countries for old men

In theory, the passing of an autocrat could create space for free elections that Mr Wine or Mr Besigye might even win. In reality, that is unlikely. Mr Museveni came to power at a time of democratic resurgence, when many other African countries ditched military dictatorship for multiparty democracy. Back then, Western donors sometimes used their influence to promote human rights.

Now, by contrast, coups occur with impunity and plenty of leaders who won power democratically are looking for undemocratic means to hang on to it. Uganda’s plight should serve as a warning to its peers: without real elections to refresh the faces at the top, expect trouble as autocrats age. ■

ch060.xhtml

In many countries, the biggest barrier to prosperity is in the home

A less sexist world would be much richer

Photograph: Getty Images

Sep 3rd 2026

All around the world, women have struggled for equal treatment in the workplace. But in some countries the struggle is more basic: to be allowed into the workplace at all. In an experiment, married mothers in Egypt were offered free or cheap child care near where they lived, to make it easier to go out and find jobs. Only 11% took up the offer. The obstacle, for many, was husbands who thought it wrong for women to work outside the home. And although 90% of Egyptian women disagree, the weight of tradition can be hard to defy—as can the literal heft of a reactionary husband.

In much of the Middle East, north Africa and South Asia, many men believe their family’s honour depends on female seclusion. They fear that if a wife goes out to work any farther away than the backyard vegetable plot, she might meet a man to whom she is not related and bring disgrace on her kin. This backward belief may be enforced with threats. In India violence against women is so normalised that nearly 40% even of women think a husband is sometimes justified in beating his wife. In Mali the figure is nearly 70%.

Such attitudes make whole populations poorer. In South Asia only a third of women are in the labour force; in the Middle East and north Africa, only a fifth are. Since 1990 the share of labour income that accrues to women has risen from 35% to 44% in liberal France; in patriarchal Pakistan, from 1.5% to a still-woeful 9%. Removing the barriers to women working would raise income per person by a fifth in many countries, estimates the World Bank—a bigger economic benefit than avoiding a typical civil war.

That is more than enough reason to care about female seclusion. But there is another. When women earn, the balance of power at home shifts. The Economist analysed data from 27 countries and found that, after correcting for how sexist a country is and how well-off a household is, women who work for pay have more say on everything from how to divvy up the family budget to whether they can go and visit a friend without asking their husband’s permission. Paid work, in short, makes families more equal.

So opening the glass front door should be a priority in all countries where it is currently shut. Alas, it has several complicated locks. The first key to insert is legal. By the World Bank’s count, 540m women live in countries where formal equality before the law is not merely imperfect but “distant”. Yet even the most sexist laws can be scrapped, as Saudi Arabia has shown. Before a series of reforms that started in 2011, women there were barred from all but a handful of jobs and not even allowed to drive. Now they are free to work, drive and shun the hijab if they choose. The share of women in the labour force has nearly doubled since 2010, from 18% to 34%. That is startling progress for a kingdom many thought hopelessly stuck in the past—even if there is still a long way to go.

Another key is social norms. These are often assumed to be intractable, yet history shows they can change rapidly. Until 1975, British banks could refuse a woman a mortgage if she lacked her husband’s consent; until 1997, Germany did not criminalise marital rape. Since then British and German cultures have evolved so that such rules are unthinkable; indeed, few young Britons or Germans are aware they ever existed.

Norms can shift with a shove from above, as in Saudi Arabia, or by a process of social contagion. Few people resolve moral questions by reasoning from first principles. Mostly, they reckon something is fine if they see lots of people they know and respect doing it. So pioneers play an essential role: the first woman in a conservative Pakistani village to brave the gossip and find work in a nearby town paves the way for others to follow. Schools, NGOs and policymakers can nudge the process along. One experiment in India found that if you show men that women they know are working in a local factory and that it is clean and not dangerous, they are more likely to consent to their own wives working. Anything that makes work safer—such as India’s women-only buses for late-night commuters—can erode the misconception that women must be kept at home for their own protection.

Digital technology can help, too. Via small screens in their hands, women in the most repressive countries can see—and therefore imagine—a different way of life. This need not mean copying the most liberal societies. More often, it means learning from the somewhat less sexist country next door, as when an Iraqi housewife sees women doing high-status jobs in Turkish soap operas, and starts to question her lot.

The last key is money. It is useful stuff, as even reactionary men agree. One economist talks of the “honour/income trade-off”: at some point, the extra cash a working wife could bring in starts to look so tempting that her husband asks whether it would really be such a terrible blow to his honour to allow it. Thus, the love of money, though much derided in scripture, can be a force for social change. ■

ch061.xhtml

What the Bayeux tapestry teaches about surviving against the odds

A 950-year-old piece of linen yields lessons for modern viewers

Photograph: La Fabrique de Patrimoines en No

Sep 3rd 2026

Days before the liberation of Paris in 1944, Heinrich Himmler, Hitler’s SS chief, sent an urgent coded message. “Do not forget to bring the Bayeux tapestry to a place of safety,” he urged the city’s Nazi commander, who surely had other things to worry about. The Nazis were obsessed with this artwork depicting the Norman conquest of England, seeing it as an icon of Aryan glory. Himmler planned to display it in Wewelsburg Castle, an SS hangout he hoped would become the “centre of the new world”. Instead, British codebreakers intercepted the order and shared it with the French resistance, who hid the tapestry in the Louvre’s basement.

Trajan had his column. Napoleon had his arch. William of Normandy’s tapestry is less imposing yet somehow more memorable. Its strip-cartoon style is jaunty, memeable and has been used to satirise everything from Brexit negotiations to Andy Burnham’s “Northern conquest” (the man he replaced as Britain’s prime minister, Sir Keir Starmer, was pictured on the cover of the New Statesman with an arrow in his eye, like the hapless King Harold). Now the tapestry is back in England for the first time since it was stitched, probably by defeated Anglo-Saxons in Canterbury, nearly 1,000 years ago. It will be on view at the British Museum from September 10th. It offers three lessons for modern viewers.

First, to survive against the odds, it is sometimes best not to attract too much attention. For several centuries the tapestry is believed to have been rolled up, stored in a chest and brought out just once a year, to be displayed in the nave of Bayeux cathedral. For about two-thirds of its existence, it was not famous at all. Only in the early 1700s did a broader audience start noticing it. Its low profile helped ensure its safety.

So did humble materials: linen and woollen thread. Many tapestries from the Middle Ages boasted gold and silver threads or were adorned with pearls and gemstones to catch the flickering candlelight. If you haven’t seen or heard of many of them, that is because they no longer survive. Their costly decoration made them vulnerable to being pinched or sold for parts. Even today, bling can bring trouble. Some of France’s crown jewels were stolen from the Louvre last year; last week a royal diamond necklace was nicked from a museum in Vienna. No thieving hands have touched the Bayeux tapestry. (Admittedly, it is hard to steal something as long as a Boeing 747.)

The second lesson is that it helps to have capable people looking out for you. When French revolutionaries eyed the tapestry to cover wagons carrying ammunition, it took the pleas of a French lawyer, and the bribe of alternative materials, to stop them. During the Franco-Prussian war, caretakers helped preserve the tapestry by keeping it locked up and hidden in a zinc cylinder. The French resistance members who saved the tapestry from the Nazis were not fighting on Normandy beaches, but they were fighting over Norman heritage.

The third lesson comes from the story the tapestry tells: it is about grace in victory. The tapestry is war propaganda, probably commissioned by Bishop Odo to glorify his half-brother’s feats at the Battle of Hastings in 1066. Yet its version of events is surprisingly sensitive to the defeated English—perhaps because there were rebellions in England in the aftermath of the battle, and it did no good to pick at the wounds of the wounded. The tapestry does not play down the horrors of war. Nor does it gloat, grave dance or call the English “losers”.

Leaders who chase attention, slather their monuments in bling, treat foreigners with contempt and constantly boast of victories they never won, should pay heed. A thousand years on, William is remembered as “the Conqueror”. What label will stick to today’s mighty rulers when they are gone? ■

ch062.xhtml

The moral panic over data centres is foolish

A backlash fuelled by misinformation and misunderstanding

Photograph: Getty Images

Sep 3rd 2026

THE SUPPLY of capital is not the only constraint on the investment of America’s AI companies. They must also worry about politics. In growing numbers and with growing fury, American voters oppose the building of data centres, which provide the computing power to train and run AI models. The opposition spans the nationalist right and the environmentalist left. Such is the backlash that the infrastructure has become one of the defining issues of this year’s midterm elections. Politicians are racing to outdo one another with restrictions, moratoriums and bans on building. Voters and politicians alike are making a big mistake.

One popular myth is that data centres guzzle more water than Agastya—an argument helped along by the grossly inaccurate calculations of a bestselling book. In reality, a mid-sized data centre uses roughly as much water as two golf courses and considerably less if it recycles its water, as many now do. And there is no reason to single out AI: just about everything in modern life, from streaming Netflix to running a dishwasher, consumes water at levels that can be portrayed as wasteful by hair-shirt environmentalists. Google has calculated that the median Gemini text prompt consumes about five drops, meaning even tens of thousands of AI queries would consume less water than a minute in the shower.

Concerns over electricity use are more numerate: data centres really do need a striking amount of power. However, there is little evidence so far that they have raised energy prices. In fact, by increasing demand they are spreading the fixed costs of infrastructure among more kilowatt-hours sold. In recent weeks a number of large utilities have linked data centres to reductions in bills. Many data centres are also helping finance new sources of generation.

Eventually the use of AI could push up electricity bills. But the answer to competing demands on resources—whether water, electricity or something else—is to price them according to their scarcity, thereby encouraging more provision and higher living standards over time. It is not to ration inputs according to the arbitrary judgment of scolds. Where pricing is difficult to calculate—to account for noise pollution, say—planning judgements are necessary, and except for needing to be near the grid there is no good reason to put data centres next to homes. In any case, even the noise worry is overstated: many viral videos of supposed AI data centre noise are really Bitcoin mines, not state-of-the-art hyperscaler facilities.

Set against the illusory costs of data centres are the tangible benefits. At a national level, data centres power a technology for which there is rapidly growing demand and on whose success future economic growth and national defence may depend. More prosaically, data centres pay huge tax bills, helping fund schools, libraries and police stations. In Loudoun County, Virginia, a tax on computer equipment inside its more than 250 centres is expected to supply 40% of the county’s total tax revenue next year. The windfall has allowed the county to cut residents’ property taxes by about 30%.

There is a respectable argument for slowing down AI research, but it rests on the danger to all humanity of the technology being abused or going rogue, not the local harms of data centres. America unilaterally making it harder to build AI infrastructure will simply hand the advantage to China, whose authoritarian leaders can put data centres wherever they choose. For years, America’s populists complained that its physical economy had been hollowed out, giving China an economic edge and leaving Uncle Sam without the industries it would need in wartime. By opposing data centres, they risk causing those exact harms. ■

ch063.xhtml

Europe cannot protect Russia from itself

Fear of Vladimir Putin going fully rogue has paralysed the West for long enough, writes Radek Sikorski

Illustration: Dan Williams

Aug 27th 2026

“THE WORLD is a healthier but not a safer place,” wrote The Economist in 1953 after the death of Joseph Stalin. This newspaper was not alone in its apprehension. Well-known Russian totalitarianism seemed somehow more manageable than whatever the future held. Fast-forward seven decades and the sentiment has changed little. No matter what the Kremlin does, fears that, if resisted, it may do something even more reckless are still used as an argument for caving in to Russian demands.

Take a conversation with Andrey Melnichenko, a Russian oligarch, published recently in these pages. He acknowledged that Vladimir Putin had made mistakes but argued that whatever comes next could be worse still; that Russia, if pushed too hard, might “descend into chaos” or even “use a tactical nuclear weapon”. The bottom line was clear: let Russia do as it pleases, or else. It is a dead-end approach.

For years after the fall of the Soviet Union, the West bent over backwards not to humiliate Russia and to turn it into a responsible international stakeholder. In 1997 alone, Moscow was invited to join the G7, signed the NATO-Russia Founding Act and entered into a Partnership and Co-operation Agreement with the European Union. In the post-Soviet era, each American presidential administration has tried to bring Russia into the fold by signing treaties, looking its leader deep in the eyes or pressing reset buttons. So did former Soviet satellite states, including Poland.

In 2009 Mr Putin was invited to attend ceremonies in Poland to commemorate the 70th anniversary of the outbreak of the second world war. My country introduced a visa-waiver programme for Russians living close to our border in the Kaliningrad exclave. The heads of the Russian Orthodox and Polish Roman Catholic churches called jointly for “forgiveness of wrongs, injustice and every evil committed against each other”. I am still personally vilified by our opposition for saying that I could imagine a genuinely democratic Russia joining NATO, once it met all the criteria.

We all tried. And we were all turned down.

Today, Russia is again trying to get its way by threatening escalation. It recently warned Britain of “consequences” for providing Ukraine with military assistance. The visit to Moscow of John Ratcliffe, the CIA director, in August was allegedly an attempt to warn the Kremlin against assaulting a NATO member state.

And yet, while the Russian leader displays no willingness to reach any peace deal—save Ukraine’s capitulation—still too many refuse to acknowledge the reality. Some argue that Russia has become so weak under mounting economic and military pressure that the West should negotiate before the regime collapses, or worse, before the state fragments and produces even more dangerous successors to Mr Putin. Others insist that Russia should not be pushed too far because it might resort to further escalation, including in the nuclear domain. Both arguments are misguided.

Fears that sustained pressure would result in the state’s break-up are overstated. Russia’s highly centralised political system has made its regions economically, administratively and financially dependent on Moscow. Millions of citizens rely on transfers from the federal budget, and even resource-rich regions depend on centrally controlled transport and energy infrastructure. Russia is not the Soviet Union. Outside the North Caucasus, ethnic Russians remain an overwhelming majority, and the memory of economic crisis and political chaos that accompanied the Soviet Union’s dissolution continues to dampen separatist sentiments. Last but not least, the country has been heavily centralised for centuries. Even in the event of a leadership crisis, or a grave military setback, the more likely consequence would be a struggle within the ruling elite to control the centre, not a dissolution.

Threats of escalation—issued either by the authorities or suggested by allegedly concerned oligarchs—are not reactions to new crises but part of a long-established playbook. Today, as in the past, Russia is testing the West by force. The Soviet Union used limited attacks to expose opponents’ strengths and weaknesses before launching a larger offensive. The Kremlin still applies the same principle. In its current fight against Western democracies, it deploys drones, cyber-attacks, sabotage, GPS interference and information warfare. The question is no longer whether Russia is probing NATO. It is whether NATO understands that it is being probed, and what signal it intends to send to prevent grey-zone actions from turning into full-fledged confrontation.

The Kremlin operates on a simple assumption: not that Russia is stronger, but that democratic adversaries will hesitate to test its vulnerabilities. Europe should therefore resist the temptation to engage in premature negotiations. Those calling to reopen political dialogue may hope that discussions will eventually reduce tension. In the long run the opposite is true. Dialogue begun before Russia’s strategic position has weakened sufficiently may strengthen the Kremlin’s domestic legitimacy, encourage further hybrid operations, undermine European support for Ukraine and deepen divisions within our democratic community.

The uncomfortable truth is that Europe must accept greater risk in the short term to avoid far more serious dangers in the future. Our objective cannot be merely to avoid escalation. It is to convince Mr Putin that escalation will get him nowhere, and that we are willing to sacrifice more than money to stand up for ourselves. European leaders’ primary responsibility is to protect their people from Russian threats, not to protect Russia from itself.

Russia regards us as the enemy and wants to destroy our democratic civilisation. We should not rush to appease it. When Mr Putin is ready for peace he will find a way to communicate it. ■

Radek Sikorski is Poland’s foreign minister and deputy prime minister.

ch064.xhtml

The world may have less time than it thinks on climate change

If it is affecting circulation patterns that determine how and where anticyclones form, it could accelerate alarmingly, warns Tim Palmer

Illustration: Dan Williams

Aug 30th 2026

THE WEATHER this summer in southern England, where I live, has been brutal: exceptional heatwaves and record-breaking spells without rain. Tracks have buckled, disrupting transportation, and crop yields are down substantially, with farmers warning of food shortages next year.

Some experts tell us that this is merely what we can expect based on long-standing climate-change predictions of “hotter drier summers”. But such predictions do not explain what we experienced this summer and it is important—and urgent—to understand why not, at least if we are to be prepared for next summer and those immediately after.

The perhaps familiar predictions of “warmer wetter winters, hotter drier summers” describe what is called the “thermodynamic” aspect of climate change: the generic consequence of the fact that global surface temperatures are warming from greenhouse-gas emissions, and that a warmer atmosphere necessarily contains more water vapour. With more humid air, the standard low-pressure systems that bring rain to the British Isles in winter will bring even more rain under climate change. In summer, by contrast, where such low-pressure systems are suppressed by the reduced equator-to-pole temperature gradient (a consequence of the Sun moving northward), the warmer air will dry the land, reducing rainfall further. We understand this thermodynamic signal extremely well—not least because the underlying physics is simple.

But none of this is directly relevant to what has happened this summer. The proximate cause of the drought was a persistent high-pressure system—an anticyclone—sitting over much of Western Europe from late spring onwards. This anticyclone steered rain-bearing weather away from Europe. To understand why it was so dry and hot this summer, we must first understand the causes for this exceptionally persistent anticyclone.

There are, broadly, two possibilities. The first is that it is part of the internal chaotic variability of climate, and nothing directly to do with climate change. If this is so, then the likelihood of a similarly persistent anticyclone in 2027 is rather small. That is, we can expect more “normal” types of circulation patterns to occur next summer. The thermodynamic element of climate change will still leave its mark even if more normal types of circulation return. The summer of 2027 can be expected to be warmer and drier compared with 20th-century summers. But we should not expect the extreme drought and persistent heat that we experienced this year.

The second, much more disturbing possibility is that the persistent anticyclone has itself been made more likely by climate change. If this is the case, then climate change is not only affecting the thermodynamics of the atmosphere but is also changing its “dynamics”. This is a crucial distinction. To understand climate change at the regional level, it is not enough to think of climate merely as some giant steam engine, warming and humidifying as we shovel in more fossil carbon; we also need to track and predict the forces and movements of the individual parts—the cogs and gears—which make up the climate system, notably the atmosphere and oceans.

In this second scenario, climate change is affecting the circulation patterns themselves, including the jet streams which determine how and where anticyclones form. There is observational evidence that persistent anticyclones are indeed becoming more frequent in summer and are thus correlated with the rise in global temperatures. Producing model data that support these observations is challenging: unlike the thermodynamic aspect, the dynamics of climate change are exceptionally complex and highly non-linear, and can only be addressed quantitatively using the most advanced climate models.

Taking these ideas to the extreme, it is possible that we may shortly pass a kind of tipping point where the dynamical effects of climate change have become so strong that persistent anticyclones will soon become overwhelmingly likely each and every summer. If this is the scenario we now face, we do not have the luxury of adapting to the new climate normal in the coming decade or two. We must somehow adapt immediately.

Some climate modellers will think I am being alarmist. After all, the most comprehensive global climate models do not support a prediction where persistent anticyclones become a feature of every summer. However, we also know that these same models struggle to simulate the type of persistent anticyclone observed this year, probably because they lack sufficient spatial and temporal resolution to represent the turbulent interactions between different scales in the atmosphere. These so-called scale interactions are known to be important in maintaining anticyclones over long periods of time, like the Great Red Spot on Jupiter. The failure to simulate this potential tipping point in current models is not a reason to be complacent but a reason to urgently improve our models.

There is an additional wrinkle to this story: El Niño is on the way. In Britain, we can expect this transient warming of the tropical Pacific ocean to have its biggest effect in winter, with the coming one wetter than normal at first, but perhaps drier than normal later. However, there is much variation in European weather from one El Niño event to another. If this one brings below-average rainfall this winter, reservoirs won’t get topped up before next summer, with potentially very serious consequences.

What to do? I have argued for many years that national climate centres around the world should pool human and computing resources to enable ultra-high-resolution global models to be built. I can hardly think of a more important and urgent issue than knowing whether we are passing major climate tipping points. If humanity works together, this issue can be resolved quickly.

But what about practical policy measures? What emergency measures could Britain introduce now to help mitigate a drought like 2026, were it to occur in 2027 (and if the intervening winter is not as wet as we would like)? Does the country need desalination plants, for example? At the least, government agencies need to think about this, urgently and collectively. There is no time to lose. ■

Tim Palmer is Royal Society Research Professor Emeritus at the Department of Physics, University of Oxford.

ch065.xhtml

Nvidia is the central bank of AI

But will its loans prove sound?

Image: Carl Godfrey

Sep 3rd 2026

It took thirty years for Nvidia, an American firm whose chips power much of the world’s artificial intelligence, to reach a valuation of $1trn. Getting to $2trn took only nine more months. It passed the $5trn mark less than two years after that. It is now the world’s most valuable company, worth around $5.4trn. Next year its sales are expected almost to double. Some analysts predict it will take in $1trn in annual revenue by 2029.

It is not just chipmaking that has spurred Nvidia’s stunning growth, however. Its boss, Jensen Huang, has also resorted to financial engineering to boost demand for its wares. In mid-August, for example, Nvidia agreed to provide a backstop worth up to $105bn for a vast data centre in Ohio that will use lots of its chips. A week earlier it had revealed plans to mobilise more than $500bn of investment in AI infrastructure with the help of six big Wall Street firms, by guaranteeing the value of the equipment it sells to such projects. It may underwrite as much as a quarter of the cost of some investments. In July Nvidia helped some customers finance big data centres in another way, by promising to top up their income if it misses set targets.

Companies often lend to their customers: think of the financing arms of carmakers, for example. Mr Huang argues that Nvidia is simply helping to unlock investment for perfectly viable projects that might otherwise struggle to borrow enough at the right price. But to critics, these deals carry more than a whiff of the dotcom boom of the late 1990s, when telecom-equipment-makers such as Cisco and Lucent lent billions of dollars to telecoms providers that bought their gear. When demand for the telecom firms’ services fell short, some of them collapsed, landing Cisco and Lucent with big losses.

As Jay Goldberg of Seaport Research Partners, a firm of analysts, puts it, Nvidia is walking a fine line between “enabling demand” and “creating it”. He does not yet think Nvidia has crossed that line, although it is “getting pretty close”. Others are more sceptical still, including Michael Burry, an investor who made a fortune betting against the mortgage-backed securities that precipitated the financial crisis of 2007-09. He and others are asking what might happen if demand for AI chips grows more slowly than expected, supply increases and prices fall, which would not only reduce Nvidia’s own profits but potentially also generate losses on its lavish support for its customers.

Lender of vast resort

The question is pertinent because Nvidia’s financial commitments are huge and growing fast. Over the past three years it has pledged over $70bn in investment in startups and offered $300bn in financial support to its customers. Some have taken to calling Nvidia the “central bank of AI”, because it plays so pivotal a role in financing the industry. That will indeed, as Mr Huang argues, help the industry grow—but it also carries risks.

Nvidia’s financial engineering is partly a response to its biggest customers’ transformation into rivals. “Hyperscalers”, tech giants such as Amazon, Google, Meta and Microsoft, account for roughly half of Nvidia’s revenue. This year they are projected to invest around $800bn, largely on AI infrastructure. But most of them have begun designing their own chips, which puts their future purchases from Nvidia in doubt.

For the hyperscalers, these custom chips are much cheaper, costing between a fifth and a third as much as Nvidia’s. They can also perform some tasks better, because the hyperscalers can tailor them to their own software. Worse, from Nvidia’s point of view, a few of the tech giants are not simply developing chips for their own use, but also marketing them to others. Google, for example, has sold some specialised processors to Anthropic, a big, free-standing AI lab. Amazon also expects its custom-chip business to become a sizeable source of revenue. Bloomberg Intelligence, a data provider, predicts that custom chips will gradually eat into Nvidia’s sales, accounting for about 50% of the market for processors used in AI by the end of the decade, up from roughly 40% this year.

Hyperscalers have investment-grade credit ratings, which keep their borrowing costs low. Upstart neoclouds have similar spending needs, but little revenue (see Business section). Their loans are naturally much more expensive. Alphabet, Google’s parent company, sold $2.75bn of 50-year bonds in November, at an annual interest rate of 5.7%. The rate at which CoreWeave, the biggest neocloud, borrowed $2.6bn in July was almost double.

It is this gap, between hyperscalers’ borrowing costs and everyone else’s, that the bank of Nvidia would like to narrow. One way it does that is by taking equity stakes in startups that will be customers themselves or that will help fuel demand for Nvidia’s chips indirectly. Last year Nvidia made about 90 such investments, nearly twice as many as two years earlier. This year it has already agreed another 60-odd.

Qualitative easing

Some of these cheques aim to propagate open-weight AI models, which users can download free of charge and adapt, unlike proprietary offerings from firms like Anthropic, OpenAI and Google, which users tend to access via subscriptions and whose inner workings are hidden. In August Nvidia agreed to pay Poolside, a startup building AI coding models, $6bn to license its software and a further $1bn for a stake. It has also reportedly agreed to buy Hugging Face, a platform hosting open-weight models, for $12.9bn. The intention behind such investments is to fuel demand for Nvidia’s chips by creating a proliferation of AI products and companies that are independent of the hyperscalers.

Increasingly, however, Nvidia is not simply investing in promising AI firms, but also helping them finance big projects more directly. In early July it announced a new stratagem in which it promises to top up neoclouds’ income from new data centres to an agreed floor. These undertakings, the exact terms of which vary from deal to deal, often last for six years. Throughout that period, Nvidia promises to pay a set price for “compute”, as the jargon has it. If the neocloud manages to sell the capacity in question at a higher price, Nvidia receives a share of the difference. This safety-net makes neoclouds’ future revenues much more predictable and so lowers the cost of the debt they take on to build new data centres. That, in turn, spurs demand for Nvidia’s processors.

Sharon AI, an Australian neocloud, expects to make use of a backstop of this sort worth about $4.9bn to deploy around 40,000 Nvidia chips. Firmus, another neocloud, plans to buy as many as 170,000. Dan Nishball, an analyst at SemiAnalysis, a consultancy, argues such arrangements need not be a permanent feature of the industry, but instead “buy time” for lenders to become more comfortable about lending to such projects.

Take CoreWeave, a neocloud firm which plans to rent out AI computing capacity.

Nvidia has invested over $2bn in CoreWeave and owns around 11% of the company.

CoreWeave buys AI chips from Nvidia to build data centres.

Nvidia agrees to buy up to $6.3bn of unused data-centre capacity from CoreWeave until 2032.

This is one of many deals Nvidia has done that use complex financial arrangements.

It has become financially entangled with many of its own customers: neocloud providers, AI labs and even investment firms.

Nvidia holds equity stakes in many of these companies.

And it has also provided financial backstops to many of its customers. In theory, these commitments could expose it to costs of nearly $300bn.

Nvidia is also taking a creative approach with big AI labs, which need huge amounts of compute but are bleeding cash. The giant data centre it is backing in Ohio is owned by SB Energy, a unit of SoftBank, a Japanese conglomerate. OpenAI will be the tenant. Nvidia, for its part, will provide guarantees for the lease on the land and buildings and for the power-purchase contract on which it will rely. In return, the project will use 1.5m Nvidia processors, with multiple “upgrade cycles” expected over the life of the project.

Even Anthropic, which buys chips from Amazon and Google, has not escaped this web. Through a convoluted series of agreements, it has reportedly signed a $35bn deal to rent cloud-computing capacity from Lambda, a neocloud in which Nvidia has a stake. Lambda, in turn, uses a facility being developed by Hut 8, another neocloud, whose entire capacity Nvidia had leased (and may now have largely sublet).

Nvidia is exploring creative ways to draw more outside capital into such projects. Its $500bn partnership with big Wall Street firms will court institutional investors such as sovereign-wealth funds, insurers and pension funds. These investors will finance independent vehicles that will buy Nvidia hardware, build infrastructure and sell compute. Nvidia will neither contribute any cash upfront for these deals nor take on any debt. In some cases, though, it will provide “residual-value support” of up to a quarter of the deal’s total price to cover shortfalls in the value of the hardware involved after a certain period. It will also provide technical support, giving lenders more confidence that the equipment they finance will be used well. Mr Huang frames this as a way to make compute “an investable asset class”.

Underpinning this web of obligations are two fundamental assumptions: that Nvidia’s chips will retain their value and that demand for compute will continue to grow at a rapid pace. Neither is assured.

Image: Carl Godfrey

Start with the chips. Mr Huang says they are “fungible” and durable, with software upgrades enhancing their usefulness for years after they are sold. They should be seen as bankable assets, he argues, that can serve as collateral for loans and guarantees. But how long an AI chip retains its value is a matter of debate. Mr Burry has argued that big cloud providers have inflated profits by depreciating them over five or six years rather than two or three.

So far Mr Huang has the better of the argument. Older chips are still in demand. In August CoreWeave announced a contract involving Nvidia’s A100 chips that runs into 2029. The A100 was launched in 2020 and Nvidia has since rolled out multiple newer models. There is evidence, too, that old chips retain their value. SemiAnalysis estimates that renting Nvidia’s H100, a workhorse of the AI industry, costs around $2.80 an hour on a one-year contract. That is only about a tenth less than when the chip launched in early 2023.

Yet both the longevity and value of such chips may simply be artefacts of scarcity. When compute is constrained, firms have no choice but to keep older chips humming. Many are using such chips for inference, meaning when an AI model responds to queries, a far less demanding task than training models, for which the newest hardware is used.

Nvidia itself releases spiffy new chips every year, with the explicit intention of superseding their predecessors. That leaves it in the odd position of arguing both that old chips should retain their value and that its new chips should supplant them. Finding a balance will be tricky.

Wayward guidance

The price of compute should fall as supply increases. Hyperscalers and even labs like OpenAI are rolling out purpose-built inference chips, since spending on inference has overtaken spending on training. Nvidia enjoys enviable gross margins, of around 75%, against roughly 55% at AMD, a smaller rival. Tim Davis, the founder of an AI hardware firm acquired by Qualcomm, a chip designer, believes that, as options proliferate, the economics of chipmaking will “start to compress”.

The biggest risk to chip prices is demand itself. As long as spending on AI keeps booming, Nvidia sells its chips, its customers fill their data centres and its various guarantees are seldom invoked. But if demand is not quite as expansive as markets expect, neoclouds and other AI providers could struggle to sell their capacity or turn a profit. That, in turn, could trigger Nvidia’s backstops, forcing it to buy unused compute, cover price shortfalls or pay for unwanted power. The same slowdown would also dent Nvidia’s sales and thus the cashflow available to meet those obligations. This scenario does not hinge on a collapse in demand; merely disappointing growth could conceivably spell trouble.

How much could Nvidia have to cough up? It has promised around $25bn in future equity investments. It owes around $33bn in debt. Its potential liabilities to customers amount to about $300bn, but only come into play in a downturn and so do not appear on its balance-sheet. These include the $105bn guarantee behind OpenAI’s data centre; as much as $125bn through the Wall Street partnership; and around $67bn in other backstops.

Nvidia will almost certainly not end up paying anything close to the full amount of its guarantees. Its commitments are spread over many years so could not come due all at once. Its proposed backstop for OpenAI’s data centre, for example, runs for 20 years, starting in 2028. If OpenAI stopped paying the lease, Nvidia could find another tenant, vastly diminishing its exposure.

What is more, Nvidia’s own finances are strong enough to weather these liabilities. Morgan Stanley, an investment bank, reckons Nvidia’s “all-in” debt will rise from $53bn early next year to $200bn by the beginning of 2029 as guarantees come into effect. But that is offset by a stash of cash and liquid securities currently worth $99bn, and a business that will generate about $200bn in cash this year. Only a cataclysmic downturn that caused all Nvidia’s guarantees to come due and its profits to evaporate almost entirely would imperil the company—as things stand.

The picture may change, however, if Nvidia’s commitments keep growing. Andy Li of CreditSights, a financial-research firm, worries that it will keep “pushing the pedal” until “something breaks”. SemiAnalysis estimates that Nvidia takes on roughly $5.9bn of guarantees for every 100 megawatts of data-centre capacity covered by its neocloud backstop programme. If it adds more such guarantees, SemiAnalysis reckons its exposure could reach $175bn by the end of 2028 from this initiative alone.

Nvidia is not alone in using ever more complex financial instruments to lubricate customer demand. AMD has offered to sell big stakes in itself to OpenAI and Meta in exchange for mammoth contracts to buy its chips. In April Google assembled a consortium of firms to help finance Anthropic’s purchase of $35bn of its chips. Broadcom, which develops the chips on Google’s behalf, agreed to cover any shortfall if Anthropic fails to pay.

Amid the race to lock in sales and loans, there is a danger that speculative projects that might otherwise struggle to find financing will get built. Should returns in the industry fall short or take longer to materialise than these convoluted deals assume, the result will be a profusion of expensive, idle chips. No company is more entangled in this web than Nvidia. Mr Huang believes that the AI infrastructure buildout is “at full steam”. But Nvidia’s investors should be aware that, the more of this boom they finance, the more of any bust they will have to bear.■

ch066.xhtml

Chandrababu Naidu’s techno-utopian vision for his Indian state

Has Andhra Pradesh’s tech-savvy chief minister bitten off more than he can chew?

Photograph: Diego Mallo

Sep 3rd 2026 | AMARAVATI

As he tells the story, Chandrababu Naidu smiles. On becoming chief minister of Andhra Pradesh, a state on India’s east coast, in 1995, he saw “IT was the future” and headed to America. Initially Bill Gates refused to meet him, but Mr Naidu insisted. In 1998 Microsoft snubbed Bangalore to establish its Indian base in Hyderabad, then the state’s capital. Today the city has 10,000-plus Microsoft employees and a big airport, and has doubled in population. More than as an investor, Mr Naidu saw Mr Gates as a signal that would propel Hyderabad’s rise.

The story captures Mr Naidu’s public persona—a single-minded tech visionary, with a knack for wooing investors and a mission to improve Indian governance. He is still at it; after a turbulent career he is, at 76, two years into his fourth term. He was almost assassinated in 2003 and briefly imprisoned in 2023; twice voters booted him out. Most painful of all, in 2014 his state was cleft in two. Hyderabad, the crown jewel, became the capital of a new state, Telangana. Andhra Pradesh, still the size of Tunisia and with a population of 54m, was left as an agrarian rump without a capital or an economic engine.

Map: The Economist

A dizzying ambition was born. As chief minister of the rump state, Mr Naidu planned a shining new capital called Amaravati, the biggest greenfield city in India’s history, and among the biggest under way anywhere. And he formulated a new pitch for Andhra Pradesh, not as a rival to India’s software hubs, but as a magnet for advanced manufacturing, data centres and quantum computing. To further these goals, he has allied his regional party to Narendra Modi’s Bharatiya Janata Party (BJP), which rules in Delhi, supporting it as a coalition partner in exchange for financial backing.

At his office in Amaravati—next to the 33,000-acre site on which 25,000 workers are busy, night and day, turning his vision into reality—Mr Naidu tells The Economist that he is building the most beautiful city in India. The Singapore-inspired master plan features abundant water and greenery and miles of cycle lanes, designed around a grid system. He will have self-driving cars on the streets within five years, he predicts. Some 300 acres have been allocated for drone research and manufacturing, and he envisages 10,000 flights a year in short order, allowing residents to enjoy streets lined with trees and free from lorries.

Mr Naidu’s vision sounds utopian, but he argues it is rooted in reality. India has long lagged behind on hardware, but it is beginning to emerge as a hub for electronics manufacturing. Mr Naidu peppers the conversation with references to Chinese prowess, which India must emulate. For him factories, data centres and quantum computers in Andhra Pradesh will complement, rather than compete with, India’s startup and outsourcing hubs in Bangalore and Hyderabad.

Yet the task is daunting. In 1995 Hyderabad had a tech industry and 5m residents. Today Mr Naidu is trying to build a high-tech economic hub on some old banana groves. What’s more, his project is controversial and his state’s finances are in bad shape. In 2019 Mr Naidu lost power to YSR Congress, a rival regional party that deals in populist welfarism. It mothballed Amaravati for five years. To win back power in 2024, Mr Naidu had to promise to keep an array of expensive schemes in place, crimping resources for investment.

Photograph: .

He has achieved impressive results. In the port of Visakhapatnam, Google broke ground in April on a 1GW AI data-centre project, much the biggest such scheme in India. The giant $15bn joint venture, involving Gautam Adani, one of India’s most powerful tycoons and a close ally of Mr Modi, is testimony to Mr Naidu’s political nous. It helps that he has made his state an easy place to do business. Steel, defence and aerospace companies are arriving, too.

In Amaravati Mr Naidu has again recruited anchor tenants—including IBM, a technology company—hoping others will follow. There are three universities and some 100,000 students living among the cranes and diggers. Already the outlines of sumptuous civic buildings rise from the marshland. Work is under way on 95 sites in all. “This is where the supercomputer will go!” shouts one excited foreman, above the din of activity.

In a country blighted by mediocre administrators, Mr Naidu’s audacity can feel like a breath of fresh air. Even so, he may have bitten off more than he can chew. One criticism is that Amaravati is an over-ambitious vanity project, and could become a financial millstone for the state. “He’s trying to punch above his weight,” says Parakala Prabhakar, a political economist. The bulk of the financing for the city is $1.6bn in long-term loans from the World Bank and the Asian Development Bank, which will be paid back by selling off land. The maths relies on land prices soaring, which in turn depends on the city attracting foreign investors, at a time when India as a whole is struggling to do this.

“These projects always look like utopias in expensive 3D renderings,” says Sarah Moser of McGill University, who has compiled a database of 300 greenfield cities worldwide. Many end up half-empty, some as ghost cities. As perspicacious as their progenitors may be, they struggle to bend economic geography to their will. Political winds change quickly. Consider King Abdullah Economic City, a vast project in the Saudi desert, gathering dust since the death of its namesake in 2015.

Amaravati’s fortunes rest on a delicate political balance. If Mr Naidu loses power, his opponents would have different priorities. For now he has the clout to push it forward because the BJP holds power nationally but is short of a parliamentary majority. That could change. Mr Naidu says he is with Mr Modi for the long haul, as they are “on the same wavelength” about development (and the opposition alliance remains “impractical”). But he admits that recent youth protests are a “concern for us all”.

A final doubt is whether Amaravati can really become “Quantum Valley”. Even as investment pours into AI, many experts think that quantum computers will have a complementary role, partly by performing some tasks more efficiently than large language models. Mr Naidu wants Amaravati to manufacture components. But the technology is years from commercialisation. And it is not clear which quantum technology will win (IBM specialises in one, which brings a risk of lock-in). Though quantum could be a good long-term bet, it is unlikely to yield investment and jobs quickly.

As with Hyderabad, Mr Naidu hopes that his truncated state will rise by gaining an early advantage in nascent technologies. In time he hopes to pass down power to his son, Nara Lokesh, now his technology minister. But he is not thinking of retiring soon. “Always,” he says, “one should be looking forward.” ■

ch067.xhtml

The Pacific Islands Forum fizzles as China flexes its muscles

Leaders stay away over debates about Taiwan’s presence

Taiwan has a Whipps round © Photograph: Getty Images

Sep 3rd 2026

AQUIRK OF the Pacific Islands Forum (PIF), an 18-member regional group, is the dress code at its leaders’ meetings. Hawaiian shirts are almost de rigueur. At this week’s meeting from August 30th to September 1st they were worn by both Pacific island leaders and by senior diplomats from America, Europe and the rest of Asia.

A less whimsical quirk is that three PIF members have diplomatic relations with Taiwan rather than China. When one of them plays host at their annual get-together, it offers the rare spectacle of Taiwan’s most senior diplomat, Lin Chia-lung, gladhanding and holding press conferences just like a regular foreign minister.

Map: The Economist

This year the leaders met in Palau, one of only 11 countries to recognise Taiwan. China’s outrage was inevitable. But how far it went to disrupt proceedings was still a surprise. Last-minute withdrawals by five leaders were a serious blow. PIF decisions are taken by consensus and only leaders can make important ones.

Suspicion quickly fell on China. Surangel Whipps, Palau’s president and a longtime critic of China’s attempts to buy influence in strategically important Pacific island states, told reporters that, weeks earlier, China had threatened unnamed consequences if Palau invited Taiwan.

The biggest no-show was Matthew Wale, the Solomon Islands’ prime minister since mid-May. The outgoing chair of the PIF, Mr Wale has been lobbying the forum to agree on a new treaty that would make it tougher for China to establish a military or police presence in PIF countries. But on the eve of the summit, three members of his cabinet resigned and filed a motion of no-confidence against him. Mr Wale rushed back to quell the parliamentary revolt. That left his proposed treaty without a champion in Palau.

Diplomats in Honiara, the capital of the Solomon Islands, say that they often see the Chinese ambassador hosting members of parliament for dinners, especially after Mr Wale became prime minister. One says that Chinese diplomats and state-owned enterprises frequently dole out cash to legislators to get them to back more China-friendly policies. But they could cite no hard evidence that the timing of the ministers’ defections was influenced by China in this case.

Some of the absences might have other explanations. Not all of the leaders are in rude health, and some are preoccupied by domestic politics. But one truant, Taneti Maamau, Kiribati’s president, threw subtlety to the wind by sending in his place his country’s former ambassador to China. All of this irked Taiwan’s Mr Lin, who told reporters that China was like a guest who behaves like a host, making demands and “ruining the atmosphere” of the party. ■

ch068.xhtml

Catastrophic floods warn Nepal of more disasters to come

It is woefully short of the resources and international help to cope with them

No home to go to © Photograph: AP Photo/Rajesh Kumar Singh

Sep 3rd 2026 | DELHI AND KATHMANDU

AWEEK AFTER it struck, it was still hard to gauge the extent of the catastrophe that hit the Himalayas on August 26th. But it is at least possible to be fairly precise about how it started: at 8.37am a collapsing glacier and ice-rock avalanche crashed down and dammed the Lhende river at the border between Nepal and Tibet. Three minutes later the dam breached, releasing a fury of water, rock, mud and ice down the Bhote Koshi and Trishuli rivers of Nepal’s Rasuwa region. Shibu Giri saw a wave raze an entire bridge. Just before it tossed up his lorry, he darted for the forest. After sleeping in a cave, he was lifted to safety by an army helicopter the next day. Many were less lucky; the flash floods swallowed some villages entirely.

Map: The Economist

A week later more than 1,000 people were known to have died and more than 4,000 were missing. They included several hundred foreigners, among them busloads of Hindu pilgrims en route to Kailash, a holy mountain in Tibet. Some 80 bridges and 40km of road were wiped out. The loss of 14 hydropower plants, some still under construction, knocked out 10% of Nepal’s power generation.

Nepal sits where the Indian and Eurasian tectonic plates collide and the hot monsoon breeze meets the Himalayas’ icy winds and shaky slopes. That makes it one of the most disaster-prone places on Earth. Its region is heating faster than the rest of Earth, its glaciers melting twice as quickly as they did in 2000. So in the years ahead more glaciers will collapse and more lakes are likely to burst their banks.

Nepal is among the world’s poorest countries, and looks horribly ill-equipped to deal with the calamities this will bring. Its prime minister, Balendra Shah, is a 36-year-old former rapper in post for a mere six months, after Nepali youths burned down parliament a year ago and kicked out their previous rulers. His government will have to find as much as $5bn to meet the costs of recovery, at a time when the wealthier nations it would once have looked to for help are cutting their foreign-aid budgets.

The United Nations’ climate fund lacks the money to honour even existing requests from poorer nations. America has pledged $3.6m in assistance—less than 3% of its contribution after a devastating earthquake hit Nepal in 2015. Besides limiting immediate humanitarian-relief efforts, the shortage of funds scuppers plans to build back better, by investing in climate-resilient roads, bridges and villages.

Himalayan nations, meanwhile, are too stingy and suspicious of one another to manage their shared home in harmony. The Hindu Kush-Himalaya has more than 63,700 glaciers but only 38 monitoring stations (only seven of which meet global standards). Pema Gyamtsho, who heads the International Centre for Integrated Mountain Development, a research and policy initiative in Kathmandu, the capital, says Nepal will have to rely on neighbours with greater resources to equip the region with better sensors and monitors and real-time data-sharing. But China tends to treat hydrologic-flow rates as a national-security secret, risking leaving its neighbours downstream blindsided.

Perhaps no early-warning system would have won Nepal the time to prepare for floods as sudden and ferocious as these. But to survive the devastation wrought in the years to come by its region’s mountains and waters, it will need more funds and more cross-border transparency. ■

ch069.xhtml

Japan’s mainstream political opposition is a shambles

The breakdown of a new alliance reflects deeper problems

Photograph: Getty Images

Sep 3rd 2026 | TOKYO

THE CENTRIST REFORM ALLIANCE (CRA) began with grand ambitions. In January two of Japan’s main opposition groups announced a merger of their parties in the powerful lower house ahead of an election, hoping to be a match for the ruling Liberal Democratic Party (LDP). Instead, the LDP triumphed and the CRA flopped, winning just 49 of the 167 seats its parties had held before the vote. On August 28th the CRA members announced they could not agree on a merger in the upper house; three days later, the CRA indicated it would disband in the lower house too, less than a year after it was formed.

The CRA has failed to solve deeper problems in Japan’s mainstream opposition. It merged the centre-left Constitutional Democratic Party (CDP), the main successor to the opposition party that held power from 2009 to 2012, and Komeito, formerly in coalition with the LDP. Both have struggled to elevate fresh leaders and to appeal to younger voters.

The shambles will mean less pressure on the LDP and more room for upstart populist parties. The LDP has dominated Japanese politics since its founding in 1955. When the Democratic Party of Japan (DPJ) swept into power in 2009, it seemed to herald the emergence of two-party politics. But the DPJ governed haphazardly. It was unlucky to be in power when a big earthquake and tsunami struck in 2011, triggering a meltdown at the Fukushima nuclear plant. The fallout has made its successor party, the CDP, radioactive.

In recent years scandals have tarred the LDP. The party lost lots of seats in elections in 2024 and 2025. Komeito ditched its long-standing coalition last October in protest at the LDP’s rightward turn under Takaichi Sanae, the new prime minister. By joining forces, the CDP and Komeito hoped to reclaim the political centre. They reportedly tried to entice two of Ms Takaichi’s predecessors from more liberal wings of the LDP to defect.

But their pitch to both potential allies and to voters fell on deaf ears. The two parties had to compromise on core principles to come together. While Ms Takaichi, the first female prime minister, projected energy, the CRA leaders, two older men, looked like dinosaurs. In the social-media era, such party elders should have “let someone younger be at the front”, laments one CDP member.

After the election, the CRA failed to clarify its message. Most voters thought the two parties should go back to operating separately. It was not surprising that the merger talks failed. No one wanted to join a sinking ship, especially not such a small one, says Nakano Koichi of Sophia University in Tokyo: “It’s not even the Titanic—it’s just small and insignificant.” ■

ch070.xhtml

India’s payments system is ten. It must start paying for itself

Processing 24bn transactions a month does not come cheap

Illustration: Lehel Kovács

Aug 30th 2026

IT SITS ATOP coconuts on Mumbai beach carts, alongside samosas in Delhi snack shops and among Navy Cuts at Kolkata cigarette stalls. It is displayed on rickshaw-drivers’ seat backs, utility companies’ bills and Amazon’s checkout page. It can be scanned to buy ten rupees- (ten cents-) worth of peanuts or a 100,000-rupee iPhone. So ubiquitous has the QR code become—and so quotidian its use—that it is easy to forget just how recent an addition it is to the Indian landscape.

The QR code is the visual signature of Unified Payments Interface, or UPI, a platform that celebrated its tenth birthday on August 25th and which has revolutionised how money moves in India. In the paleolithic era—before 2016—cash was used for almost all transactions by number and two-thirds by value. Cheques accounted for the majority of non-cash dealings. And for some reason everybody pretended they had no small change, giving rise to an alt-currency of one-rupee boiled sweets.

In the first 12 months after UPI’s launch, Indians used the platform to make 73m transactions worth 227bn rupees. This year, in the month of August alone, it was used 25bn times for transfers worth nearly 30trn rupees. More than 85% of non-cash transactions now go through UPI (cheques are down to just 0.2%). And there is still room for growth. The network’s 550m-odd users make up just over half of India’s adults.

There are many reasons for UPI’s success. Chief among them is its simplicity. All a user has to do to both send and receive money is to link their bank account to a mobile app. After that, making a payment is as easy as scanning someone’s QR code or entering their phone number—no need to share bank details, fiddle with plastic cards or ever go near a point-of-sale device.

Other factors contributed. Jio, a telecom network, was launched the same year as UPI. Its aggressive pricing caused mobile-data costs to plummet, bringing millions of Indians online—and thus to UPI apps. Narendra Modi’s surprise decision to invalidate 86% of cash by value in November that year—cancelling 500- and 1,000-rupee notes, ostensibly to punish undeclared cash holdings—also played a part. The policy wrecked the informal economy and slowed GDP growth, but predictably cashlessness soared. UPI transactions rose from 100,000 in October 2016 to 2m that December.

Another helpful, less damaging, policy was the government’s decision to ban fees for commercial UPI transactions. A merchant accepting a credit or debit card typically pays for the privilege. That fee is split between the network provider (such as Visa or Mastercard), the card-issuing bank and others. But since 2020 a merchant has paid nothing to accept payments via UPI. That spurs adoption, too.

Shortly before UPI’s tenth birthday, parliament passed legislation unbanning those fees. Though new rates have not yet been set, it is expected that they will apply only to commercial transactions and above a minimum amount that would exclude most transfers, including to the coconut vendor, the rickshaw driver or between individuals. The idea is that only big businesses will pay. Oddly for India, where big business is unpopular, the reaction has been mostly negative. Critics of the proposed fees have argued that the government’s costs are offset by savings from printing less money; that UPI’s benefits outweigh its costs; and that consumers will ultimately bear the burden.

That last one at least sounds as if it has merit. But UPI costs money to operate. The system is overseen by the National Payments Corporation of India, a non-profit body jointly owned by public and private financial institutions. Banks spend huge amounts to maintain their networks, guard against fraud and ensure payments go through reliably. In return, the government gives them subsidies that cover only a small fraction of their costs. Without proper funding banks have little incentive to invest in innovation. Worse, they are barely able to keep up with basic maintenance and anti-fraud measures. Consumers already pay—but indirectly through taxes and a service that is not as good as it could be.

India’s government deserves praise for building UPI, for turning it into the backbone of the country’s payments and for making the right decisions about fees twice over. But none of those achievements will amount to much if the system stagnates or weakens. For years UPI has been a way for others to pay. It needs to pay its own way, too. ■

ch071.xhtml

Social-media curbs for kids are not working in Asia

They are widely ignored, flouted and sidestepped

…and pretend the grown-ups know what they’re up to © Photograph: Getty Images

Sep 3rd 2026

IT SEEMED A parent’s dream. To manage the time her 14-year-old son spent online, Harini Naidu had tried various tactics: from restricting Wi-Fi access to turning the internet off at night. So she was excited when Australia became the first country in the world to ban social-media accounts for under-16s. But the rules, imposed in December, “literally did nothing”. Her son lost his YouTube account, but can still watch videos without logging in. Ms Naidu’s daughter Ishana was 15, but has kept her accounts on X, Instagram, Snapchat and TikTok. Ishana says: “It’s so underwhelming that people have forgotten there’s a ban.” They do not even have to hide behind virtual private networks.

The experiences of teens like Ishana matter to other countries, too. Since Australia brought in its curbs, dozens have done the same, or are considering it. On August 24th New Zealand announced a planned ban. Policymakers worry that social media cause various harms to the young. They also know that bans are popular with adults. Many cheered on August 26th when Meta agreed to change its products for under-18s in most American states, including a two-hours-per-day limit on their scrolling, after settling a lawsuit alleging its platforms harm youngsters.

Yet so far, Australia’s policy has proved largely ineffective. At first the government was able to tout some successes. Social-media platforms had removed, restricted or deactivated almost 5m accounts in the first month. And it may be too early to write off the effectiveness of Australia’s new rules. But studies have shown they have so far had minimal impact on usage. One in the British Medical Journal in June found that, three months after restrictions were introduced, 85% of under-16s said they still used banned platforms. Twelve- and 13-year-olds spent the same amount of time on social media as before the ban. Worries that children migrate to alternative platforms, with lower safety standards, were mostly unfounded—but largely because so many remain on existing ones.

Young Aussies are keen to stay online. Some spy a challenge, says Ishana, as under-16s “want to use social media more because it’s banned”. The young “know how to outsmart the system”, she explains, including by creating joint accounts with friends, which she thinks confuses age-checking systems. Yet she says several accounts with her real date of birth were not removed. Many under-16s have made new accounts using a fake age. Others have entered their parents’ or siblings’ details.

Some blame tech firms for not enforcing the rules. KJR, a software-testing company, recently opened 50 social-media accounts in Australia claiming the user to be 16. It said only one site did any verification and all 50 accounts remained active. The government is trying to beef up its rules, including by doubling fines on offending firms to A$99m ($71m). It also wants to give Australia’s eSafety Commissioner power to make social-media firms hand over internal-usage data.

Others blame the policy’s ineffectiveness so far partly on the technology meant to enforce it. Firms are using various tools, including AI, to analyse posts and scan faces. But facial-age estimation software is not very accurate in distinguishing between, say, a 15- and 16-year-old. Lucina, a 15-year-old from the state of Victoria, says her face was scanned on Snapchat and she was told she looked 18. “The policy hasn’t been tested because it’s been technically impossible to enforce,” says Tama Leaver of Curtin University in Perth.

Mr Leaver thinks other countries planning similar laws are concluding that the “only way to do this is to require people’s real IDs”—ie, an official digital or physical identity card—when they sign up to social media. Malaysia, one of just a few other countries where Australian-style restrictions are in force, has opted for this method. L.S. Leonard, the chief legal counsel of Malaysia’s Communications and Multimedia Commission, the regulator, says that its assessment indicated that this would be “much more accurate” than the age-assurance approach used in Australia.

Yet things are not going to plan in Malaysia either. Mr Leonard says social-media platforms claim to have removed lots of under-age accounts, according to their own guidelines. But nothing has changed for most users, says Tricia Yeoh of the University of Nottingham’s Malaysian branch. Mr Leonard says platforms have asked for time to “make their system adaptable” for age verification; he expects everything “to be resolved” by January, after which non-compliers may be fined. Ms Yeoh, however, suspects platforms are worried about the risks of storing personal IDs on systems that could be hacked, or subject to requests from governments.

Some oppose blunt social-media bans like Australia’s because they think they let platforms get away with not making their sites safe. Ways to incentivise firms to do so are being contemplated. Canada plans to ban platforms from allowing under-16s to have social-media accounts, but only until these apps meet a set of safety criteria. And there are hopes that Meta, after introducing limits for teens in most American states, will roll the same features out elsewhere, and that other social-media firms will follow suit.

Young Australians think they deserve more credit: Ms Naidu’s 14-year-old son says many teenagers already “know how to be safe online”. He stays off X, for example, because it is “very unfiltered”. He prefers Reddit, which is “less unhinged”. ■

ch072.xhtml

Why can’t a tech-savvy China make its food safe?

Tainted vegetables are the focus of the latest scandal

Photograph: Xinhua News Agency/eyevine

Sep 3rd 2026 | Kangbao

THIS TIME it is the cabbages that Chinese shoppers are avoiding. On August 22nd a blogger posted videos of truckers in Kangbao, a county a few hours’ drive from Beijing, dipping the vegetables in formaldehyde. The chemical keeps their pale green leaves looking fresh on their journeys through the summer heat. But formaldehyde can also cause cancer. Amid national outrage, authorities have set up checkpoints to test cabbages leaving the county. So far three people have been detained on criminal charges. Singapore and South Korea have also stepped up inspections of Chinese cabbage imports.

State-run media have blasted the truckers for “breaching the moral bottom line”. Yet many Chinese are wearily unsurprised. One farm worker in Kangbao admitted he had known about the problem for some time. Reports of formaldehyde misuse go back at least as far as 2012. And the chemical is just one of the nasty substances that still end up in the country’s food. Last week a cleaning company was put under investigation for using a banned pesticide in restaurants. Bayberry-sellers were recently caught soaking their wares in another prohibited chemical to keep them fresh.

Few things anger China’s public as much as food-safety scandals. Many are still haunted by a case in 2008 when tainted milk powder made some 300,000 children ill and killed six. In 2013 Xi Jinping, China’s leader, tied the legitimacy of the Communist Party to the problem. “If our party cannot even ensure food safety,” he said, “then people will start to question whether we are qualified to govern.” Yet the scandals continue. And as China grows more powerful and sophisticated, the problem becomes more embarrassing. People can buy cutting-edge homegrown robots, electric cars and smartphones, but they still do not trust their country’s food.

Some progress has been made. China has tightened the comprehensive law on food safety it passed in 2009. Current regulations are not that much weaker than Europe’s. Punishments, meanwhile, can be far more severe: two men involved in the milk-powder scandal were executed. China has also made progress on pesticides. Farmers now use less of them, thanks to training, crop testing and bans on the most toxic substances, notes Even Pay of Trivium China, a research firm. That means fewer harmful residues on someone’s plate. One study of Sichuan province, in south-western China, found pesticides per acre had fallen by a third from 2010 to 2020.

Yet the scale of the challenge is still daunting. Ensuring food safety is tricky anywhere: this summer America is dealing with a nasty parasite outbreak spreading via lettuce. In China the difficulty is compounded by the fragmentation of the food industry. A few big agricultural firms jostle alongside 160m independent farmers, who mostly eke out a living on small patches of land in poorer parts of the country. Then there are the hundreds of thousands of companies moving grains, fruit and vegetables from farms to cities. Most Chinese people also buy their groceries from small neighbourhood markets rather than big supermarket chains. Many of them dine out in hole-in-the-wall restaurants. That all creates plenty of room for dodgy actors, operating on margins thin enough to encourage taking short cuts.

Regulatory bodies are as fragmented as the industry they govern. Things can fall through the cracks. In 2024 it came to light that cooking oil was being transported by lorry and by ship in tanks that were also used for fuel, without cleaning in between. Stopping such dirty tricks might have been the responsibility of China’s market regulator. But the obligation might instead have fallen to the transport ministry, or an office in charge of managing strategic resources like oil. In the end, all failed. It was an investigation by state-run media that discovered the problem.

China is hoping technology can help. Some fixes are simple: many Chinese restaurants now live-stream their kitchens, so customers can watch their food being made. Others are fancier: some big companies use blockchain, a kind of secure shared database, to track their products. China is also building out its “cold chain”: the network of refrigerated lorries and warehouses that keep fruit and vegetables fresh on their journeys without the need for, say, formaldehyde. A recent government plan called for another 500 cold-chain logistics centres over the next five years, on top of the 1,000 already in place.

Such things all add to costs, though. Wealthier Chinese seem happy to shell out. Sam’s Club, a members-only supermarket subsidiary of Walmart, an American chain, has expanded rapidly in China on a promise of supplying safe food. Hema, a homegrown alternative set up by Alibaba, a tech giant, allows consumers to scan some product barcodes to see chemical-testing certificates and transport logs. But not everyone can afford to pay much more for their groceries.

A cabbage-patchy response

To make food that is both pleasingly cheap and reliably safe, China will have to deal with deep-rooted problems. One is corporate culture. Chinese companies often pride themselves on their speed to market. But that can also entail a rushed approach to vetting suppliers and other safety precautions. Internal channels for reporting problems to management are often underdeveloped. The government is encouraging employees to report to officials instead. In 2024 it announced bigger rewards for whistleblowers in the food industry, promising up to 5% of the value of fines imposed in serious cases.

But official incentives are also warped, warns one Chinese expert on food safety. To get promoted, grassroots officials have to nurture local industries and protect local jobs. Both goals could be jeopardised by an overzealous focus on policing food. “Everyone is maintaining a tacit balance between safety and economic growth,” he says. Big scandals typically lead to investigations from higher levels of government, which will then unerringly find some local scapegoats to punish. Mindful of this, local officials often prefer to deal with problems themselves or, if possible, ignore them.

In Kangbao, where GDP per person is about a third of Beijing’s, farming represents a substantial share of the economy. Local officials are probably bracing for the fallout from the scandal. They are certainly keeping close tabs on the media. Shortly after this correspondent arrived in the county, four officers appeared in a police car, followed by another clutch of officials from the propaganda bureau. They were eager to talk about the fine weather and a rare species of bird that breeds nearby. They did not, however, want to answer any questions about cabbages. ■

ch073.xhtml

The lengthening reach of China’s armed forces

Why Chinese jets are staging mock combat over Egypt

Photograph: Getty Images

Sep 1st 2026

WHEN SIX Chinese fighter jets flew over Egypt in the last week of August, the manoeuvres on display were both aerial and diplomatic. The J-16 jets, among the longest-range and most heavily armed in China’s air force, have been doing joint exercises with Egypt for the first time. That is noteworthy, given Egypt’s close defence ties to America and new doubts about America’s reliability in the region. What is more, China’s aircraft were the first to stage mock combat with French-made Rafale fighters, which are used by Egypt but also by India, whose border dispute with China flared into deadly clashes in 2020.

The fortnight-long “Eagles of Civilisation” exercises are expected to end in early September, just as Xi Jinping, China’s leader, arrives in Egypt for his first visit in a decade. The drills and the visit were probably planned well before America’s war with Iran started in February. The aerial drills were happening for the second year running. And the official purpose of Mr Xi’s visit is to mark the 70th anniversary of diplomatic ties between China and Egypt. Nonetheless, they send a clear signal about Chinese ambitions in the region.

China is demonstrating its potential as an alternative supplier as well as its growing capacity to project air power far from home. And it is doing so just as America’s dominant military role in the Middle East is undermined by its failure to defeat Iran, the depletion of its weapons stockpiles and the damage done to its bases, which it has been unable to protect adequately.

China has neither the capability nor, it seems, the desire to supplant America in the Middle East. But it is getting steadily better at carrying out limited operations to help partners and protect its interests in a region that is vital for China’s trade and energy supplies. At the same time, China is gaining experience and data that could help its pilots in future combat against Rafales and other Western equivalents. India will be especially concerned. During its conflict with Pakistan in 2025, at least one Indian Rafale appears to have been shot down by Pakistan’s Chinese-made fighters. (Indian officials, while acknowledging that they lost some fighters, deny that a Rafale was among them.)

Egypt is not the only target of Chinese military diplomacy in the region. China has used arms sales and joint exercises to improve defence ties with other countries too, including Saudi Arabia and the United Arab Emirates (UAE), even though both have close military links to America. China and the UAE were due to start their latest round of joint air drills in north-western China in late August. These have happened yearly since 2023. Neither side has said which aircraft will take part but in the past they have involved the UAE’s French-made Mirage fighters, which Taiwan also uses.

Such drills are partly marketing exercises. In 2024 Egypt was reported to have placed its first order for Chinese fighter jets, an undisclosed number of single-engine J-10s. China denied that. But it sent J-10s to the first round of joint air drills with Egypt in 2025 in an apparent bid to clinch a sale. Egypt may have been considering them as replacements for its ageing American F-16s, which it started buying in 1980. Now China appears to be touting the heavier, twin-engined J-16 as a way to help Egypt carry out longer-range operations over the Red Sea, the Horn of Africa and elsewhere.

The exercises also give China a chance to test its ability to quickly project air power far from its own borders. The J-16s took about nine hours to fly the 6,000km from China to Egypt. They were refuelled by a Chinese YY-20 air tanker that flew with them. The Chinese fighters and other aircraft also stopped over in the Pakistani city of Karachi. Pakistan is one of China’s closest military partners, so the layover highlights its potential as a staging point for Chinese operations farther afield.

But it was the mock combat with Egypt’s Rafales that grabbed most attention back in China. Its state television broadcast footage from the cockpit of one of the J-16s showing a Rafale manoeuvring nearby. It said the drills included one-on-one and two-on-two combat. Such “dissimilar air combat training” offers pilots more valuable experience than drills with the same kind of fighters. They also let each side gather data on the other’s aircraft. How long it will be before China can engage in real combat so far from its borders is unclear. But as a future capability, it has clearly locked on to that target. ■

ch074.xhtml

The serious, rivalrous business of Chinese barbecue

Some cities say they’ve been eating kebabs since the Han dynasty

Queuers for skewers © Photograph: Getty Images

Sep 3rd 2026 | QIQIHAR

STEP ASIDE, Texas. The Chinese city of Qiqihar is laying claim to the global barbecue crown. Located in the country’s far north-east, it recently concluded a barbecue festival, inviting the world to enjoy its flavour. Banners across the city of 4m assert it is the “international barbecue capital”. That is a big claim for a place few outside China have heard of and equally few could pronounce (roughly, chee-chee-har).

Map: The Economist

Qiqihar’s real beef is not with Texas brisket or Aussie snags but with other Chinese cities. Barbecue has become a big business in recent years. The grilling industry nationwide was worth around 270bn yuan ($40bn) in 2025. Several cities are vying for the thickest cut of that total. Qiqiharians make a good case. Their region has some of China’s largest grazing pastures. In the summer locals sit on the streets and toss tasty beef on coal-fired grills along with onions and sauerkraut. They are adamant that their meat is the best in China and that they have the figures to prove it. Between 2020 and 2025 the city’s beef sales grew from around 5bn yuan to 30bn.

The biggest challenge to Qiqihar has come from Zibo, a city in eastern China specialising in meat skewers. In 2023 its barbecue stalls became a hit travel destination. Tourism income grew by 70% compared with the year before. Local officials were quick to support the trend. Banks were encouraged to hand out “barbecue loans” to help budding entrepreneurs source mutton and set up stalls. Accommodation became so scarce the government imposed price controls.

The craze in Zibo ignited a bigger barbecue battle. Jinzhou, another north-eastern town, has claimed for years to be China’s pork-skewer capital. When its officials saw Zibo’s success, they sent a delegation to figure out how best to bank their own shanks. Residents in Xuzhou, another eastern city, believed their meat-grilling tradition to be much older than Zibo’s and more deserving of tourist spending. Stone carvings from 2,000 years ago purportedly show people eating meat on sticks. Local netizens claimed Zibo had pilfered their heritage.

China’s summer travel spots often go viral on social media but quickly fade. Zibo has struggled to maintain excitement over its delicacies. Its heavy investments in grilling have reportedly led to many failed vendors. Meanwhile Qiqihar now supplies a network of restaurants across the country, which sell the region’s beef and thus support its economy. Qiqihar may not pose a threat to Texan pitmasters but it has a healthy claim on China’s cookout capital. ■

ch075.xhtml

In film and in life, China pursues dragon-restaurant diplomacy

It is full of contradictions but welcomed by many as an alternative to Western power

Illustration: Cornelia Li

Aug 31st 2026

“WELCOME TO Dragon Restaurant”, a new Chinese box-office hit set in a thinly disguised Iraq, will not win any Oscars. The plot is outlandish, many characters cartoonish. Yet people outside China should still watch it. For one thing, it depicts the American invasion of 2003 from an unfamiliar vantage: not the green-tinged light show beamed into Western homes, but the shriek and crash of bombs slamming into a slumbering city, shocking and awful.

In following a Chinese chef who profits from the war until it turns on him, the film also carries political significance. It sheds light on China’s self-image and its positioning in the Middle East and beyond. If it seems absurd to read so much into a movie, remember that commentators have used another one, “Wolf Warrior”, as shorthand for China’s aggressive diplomatic stance over the past decade. Dragon-restaurant diplomacy reflects a different style—and a better approximation of Chinese foreign policy in the round. Its organising principle is to be friends to all people, motivated more by mercantilism than political ideology. The approach, both in the film and on the ground, is often superficial and contradictory. Yet it would be a mistake to ignore its appeal to countries tired of overweening Westerners.

The hero of the film (which in English is officially called “Once Upon a Time in the Middle East”) is Xu Fu. He moves to a fictionalised Baghdad in the hope of making more money from his cooking. When war starts, he discovers the lucrative “Green Zone” that houses the American forces. Western soldiers and, ahem, journalists like his food and love his booze. He rakes it in, until a suicide-bomber strikes. His former business partner is implicated, so Xu spends time in an American jail. Later, Islamic militants capture him. Chinese state media and intellectuals have heralded the film as a parable of what China offers a world wrecked by America. Yet there is more to the film, not all of it flattering to China.

A central conceit is that China wants to be liked by everyone. Xu Fu is a paragon for any good, nimble Chinese diplomat. His local partner is an official in the pre-war regime; Xu stays loyal to him even after he loses power. American soldiers also grow fond of him and, especially, his cooking. So does the terrorist mastermind who seizes him. Through it all, Xu sees their common humanity. The film, to its credit, is not blind to the tensions. “Your partner was the perpetrator. Your friend was the victim. You really think you’ve got nothing to do with this war?” an American soldier asks him. That might prompt self-reflection in Chinese officials who protest neutrality in the Russia-Ukraine war and are peeved when Westerners point out that China’s supplies and cash keep the Russian forces going. You really think you’ve got nothing to do with this war?

Xu’s motivation also rings true. He is both candid and shameless at the outset in his desire to make a quick buck to pay off debts. As his Chinese partner reflects, profiting from war is a stroke of luck. It is a small example of the opportunistic commercialism that drives much Chinese activity overseas. It sounds grubby. But for countries on the receiving end, it is often welcome—a contrast to the perception that Western funds come with strings attached. Ren Zhengfei, founder of Huawei, a technology giant, has made it his maxim: “In business, speak only of business.” Tuning out politics is partly why China manages to straddle antagonistic divides in the Middle East and beyond. Think of Iran versus the Gulf monarchies, Algeria versus Morocco, Egypt versus Ethiopia—most countries pick a side in these squabbles. Not China. In many countries, those doing business with it happily go along with the idea, however illusory, that politics takes a back seat to business.

Another characteristic of dragon-restaurant diplomacy is its thinness. Later in the film, Xu is imbued with humanitarian aims, working to save orphans. Yet he is a reluctant saviour. At one point a friend tells him: “You should have stayed out of it. These are not your children.” In the end, he takes this counsel, returning to China and leaving it all behind. The film’s anti-war message, in Xu’s words, is to “eat well and raise your children well”. Not much to disagree with there. But it speaks to the limits of Chinese support for partners abroad. China is wary of entanglements, especially of the military variety. Iran and Venezuela learned that this year. “It is a matter of managing expectations,” says Karim Alwadi, an expert in Sino-Arab relations. “China is not interested in being a security guarantor.”

The chef’s special

For Westerners who were disturbed by China’s wolf-warrior turn, the rise of dragon-restaurant diplomacy may seem a big improvement. Sprawling across so much ground, it may appear toothless, even naive. In reality, it is a formidable challenge. For all its contradictions, it appeals more to other countries than trying to tariff or, worse, bomb them into submission. And investment dollars, when they materialise, beat morality lectures.

Yet China risks being seduced by its own cooking. One real-life inspiration for the film was Liu Lei, who quit his tech job and moved to Baghdad in 2003 to open a restaurant. In his co-authored memoir, “Making Money at Death’s Feet”, it was a young Japanese anti-war campaigner, not Mr Liu, who was the real humanitarian. She worked as a waitress in his restaurant and every afternoon collected leftovers to bring to an orphanage. That plotline would have struggled to get past a Chinese censor. In trying to be friends to all, China makes one glaring exception: Japan. Xi Jinping accuses the country of reviving its militarism. He cannot accept that Japan might want peace as much as China professes to. In dragon-restaurant diplomacy, China decides which truths to serve. ■

ch076.xhtml

How data centres became one of America’s hottest political issues

Growing hatred of the facilities is reshaping the midterm elections

Photograph: Getty Images

Sep 2nd 2026 | CHICAGO AND WASHINGTON, DC

ASPECTRE IS haunting America—the spectre of data centres. The computer-filled warehouses powering America’s boom in artificial intelligence have become less popular than almost every other kind of infrastructure—even nuclear-power plants. Counties and states are rushing to wrap them in red tape or ban them altogether. In July the Ratepayer Protection Act, designed to limit their impact on electricity prices, sailed through the committee stage in the House of Representatives.

On August 31st Donald Trump weighed into the fractious debate, alarming Republicans facing tight midterm races. Local opponents of data centres would “end up being backwards and poor”, he warned, if they killed the “Golden Goose”. The backlash against data centres is scrambling traditional political alignments, reshaping the coming elections and unnerving tech firms and investors. It is less clear whether it will slow down the ai juggernaut.

For a sense of how quickly the politics of data centres has changed, consider three very different states: Illinois, Texas and Pennsylvania. In 2019 J.B. Pritzker, the governor of Illinois, signed a bipartisan law creating incentives for their construction. “In today’s world,” he declared, “data centres are as critical a part of our infrastructure as our roads, trains and schools.” In Texas, Greg Abbott boasted as recently as last year that the state was “the epicentre of AI development” while welcoming new Google data centres. Josh Shapiro, Pennsylvania’s governor, was similarly enthusiastic: “We are already all in on AI.”

In August Mr Pritzker had a rather different message. Data centres, he told reporters, risked “higher electric rates, higher water rates or…environmentally unsound practices”. In June he had unilaterally paused applications for the tax credits introduced in 2019. In Texas, Mr Abbott ordered the state’s power utilities to audit all data centres seeking to draw electricity. The facilities “could endanger the reliability and stability of the Texas electric grid”, he warned. Mr Shapiro has also tightened the rules, giving communities more power to veto projects, while tarring his opponent in the gubernatorial race as pro-data centre—increasingly, a political term of abuse.

Chart: The Economist

A little over a year ago some polls showed majority support for local data-centre construction. By March that had flipped: one Gallup poll found 71% opposed. Surveys by Annenberg recorded a 19-point swing in net favourability between February and June-July. An Economist/YouGov poll from the end of August put opposition at 63%. Data centres were viewed as bad for the country across every group of voters (see chart 1).

Chicago illustrates the shift. Long before the AI boom, the city and its suburbs hosted scores of data centres, thanks to their position at the heart of America’s internet infrastructure and relatively cheap electricity. Until recently, few people cared. Now they are packing public meetings in an effort to stop development across the state. In August hundreds turned out to oppose plans for a project on Chicago’s South Side that, according to its developers, is not even a data centre.

Chart: The Economist

In the first eight months of this year, American counties considered 409 moratoriums or similar measures, nearly seven times last year’s total. In more than half a dozen states—including virtually every midterm battleground—more than a fifth of counties have passed at least one moratorium (see chart 2). “This is one of the toughest development issues I’ve encountered in my career from a land-use attorney’s perspective,” says Tommy Mann of Winstead, a Dallas law firm. “The level of emotion associated with these projects is one that I have not experienced.”

Some proponents of data centres have blamed foreign powers for the backlash. “China could not be happier,” observed Mr Trump. In June OpenAI, a model-maker, said that a Chinese firm had produced images designed to undermine support for data centres in America. On August 27th X, the social-media platform, said it had identified a 200,000-strong Chinese bot farm circulating the material. Yet the evidence suggests these efforts have had little effect. OpenAI concluded that they generated “little or no observable engagement” among Americans.

Too close to home

The real cause of the outrage is simpler. Americans are sceptical of AI, distrust the firms that build it and fear it will lead to job losses. But poll after poll suggests that what worries them most about data centres is their local impact—or, at least, what they imagine that impact to be. A recent Fox News poll found that 75% of respondents cited concerns such as energy use, water bills and traffic, as the main reason they opposed a data centre in their area. Just 11% pointed to AI itself.

Many of these concerns are overblown. Google, a hyperscaler, says that in 2025 its data centres and offices consumed around 10.9bn gallons (41bn litres) of water—roughly equivalent, it noted, to the annual irrigation of 73 golf courses in the south-west. Amazon, another hyperscaler, says its global fleet of data centres used 2.5bn gallons last year, down by 2% from 2024 despite adding more facilities. In 2023 crop irrigation consumed 74 times more water than data centres, even after accounting for the water used to generate their power.

Electricity is a different matter. The Lawrence Berkeley National Laboratory, a federally funded research centre, estimates that data centres consumed 4.4% of America’s electricity in 2023 and could account for between 9.5% and 15.3% by 2030. Their impact on consumers is less clear. Wholesale prices are being pushed up. That could eventually impact households. But the Electric Power Research Institute, a think-tank, estimates that the growth of data centres actually reduced average electricity prices by 6% between 2019 and 2024 by spreading the grid’s fixed costs.

In recent weeks three big utilities—Georgia Power, Pacific Gas & Electric and Indiana Michigan Power—have all linked data centres to lower bills. Meanwhile, 30 states have approved or proposed “large-load” tariffs designed to ensure that big customers such as data centres pay for infrastructure upgrades.

The question is whether public opinion is still malleable. A poll this month by Echelon Insights found that opposition to data centres could be reversed if respondents were told about just one potential benefit, such as lower property taxes or more spending on schools. “Voters are open to making a deal on data centres,” argues Patrick Ruffini, a Republican pollster who runs Echelon. They also seem open to bribes. Another poll, by the University of Virginia, found that most people would accept a few thousand dollars to support a local data centre, with a median acceptable payment of $10,000. Paying every resident of a median-sized American county would cost around $240m—pocket change for the tech industry. It is already showering gifts on communities.

Still, the backlash shows few signs of abating. “We’re going to see the same thing happen in 2028,” says Darrell West of the Brookings Institution, a think-tank. Governors who escape the issue this year will face it in two years’ time, he argues, when they come up for re-election.

The politics of data centres is topsy-turvy in other ways, too. The backlash could prove especially awkward for Democrats in Illinois, where trade unions are powerful. Many unions are enthusiastic supporters of data centres. After Mr Pritzker’s decision in June, a group of union leaders issued a critical statement. The move, they said, “will send billions of dollars in investment and thousands of union jobs to Indiana, Kentucky and Ohio”. Tim Gillooley, a local union official, is even more emphatic: “These data centres are going to be built! Let’s get them built here.”

The path of least resistance

The growing opposition is sand in the wheels of the industry. Sales-tax exemptions for computer chips—the biggest cost of building a new data centre—remain in place in most states, including Texas. But they have been paused in others, including Arizona and Ohio, or made conditional, as in Pennsylvania. That, along with satisfying local demands, will inevitably increase the cost of new data centres. Even so, a handful of states still wield outsize influence. Texas alone has more data centres under construction than Georgia, Arizona and Pennsylvania combined (see chart 3).

Chart: The Economist

Data-centre investment may simply flow to more permissive states. “Enough of these data centres are going to find a home in a community willing to work with them in return for the tax revenue and other benefits,” says Jack Painter, an energy analyst at Capstone, a strategy firm. As Mr Trump put it: “Let Data Reign.” ■

ch077.xhtml

Quantifying anger in American politics

The country is awash in indignation

Photograph: Getty Images

Aug 30th 2026

IN JUST THE past two months Donald Trump has described Democratic candidates as “jihadists” and “radical lunatics”. He has mocked opponents’ weight and appearance, said he hated Adam Schiff, a Democratic senator from California, and referred to James Talarico, the Democratic Senate candidate in Texas, as “Tal-a-freak-o”. He has also attacked former Republican allies, calling them “losers”.

Though more hostile than most, the president is hardly the first politician to trade in outrage. New research suggests he is part of a broader trend: American politics is angrier than ever.

Economists at Harvard, HEC Paris, Paris-Saclay and the University of California, Berkeley, set out to explore the evolution of political fury. They trawled more than 150 years of congressional speeches and millions of social-media posts, categorising their emotional character using artificial intelligence, including large language models and natural-language processing. For each text they assigned a score for eight emotions—anger, sadness, pride, gratitude, fear, disgust, joy and hope—as well as for the overall level of emotion.

Chart: The Economist

Their paper finds that angry statements in Congress shot up during Newt Gingrich’s leadership in the 1990s, before receding in the early years of this century. More recently such statements have become ever more common, rising by 50% since 2013 (see chart). Voters are more irate, too. From 2013 to 2025 the share of policy-related sentences expressing anger in voters’ tweets rose by 35%. In comments on Reddit, an internet forum, it rose by 46%.

The paper’s authors are careful not to attribute the rise in anger to any single cause, though there are plenty of theories. Steven Webster of Indiana University emphasises deepening partisanship. According to the Pew Research Centre, since 2016 there has been a sharp rise in the share of Democrats and Republicans who not only oppose one another’s ideas but also view members of the other party as immoral and dishonest. The share of partisans who say that the other party makes them “angry” climbed by almost 30 percentage points, to 76%, between 2016 and 2025.

“Regardless of where it starts, it is going to be a reinforcing feedback loop,” says Stefanie Stantcheva of Harvard University, who co-wrote the new research. She points to data showing that angry tweets by members of Congress tend to receive 59% more retweets than emotionally neutral ones. Anger can mobilise voters more effectively than other feelings, setting in motion a corrosive cycle. Political scientists have shown how rage undermines trust in government, reinforces polarisation and reduces voters’ commitment to democratic norms, even when the source of anger is detached from politics itself.

Dr Stantcheva and her colleagues found that respondents primed with negative emotions were less favourable towards immigration and trade, and more favourable towards economic redistribution—perhaps a sign of rage at the rich. “When you’re angry, the categories you’re retrieving are grievance and blame,” she says.

As politicians vie for attention ahead of the midterm elections, many are making anger a central part of their campaigns. In April the news site Axios reported that nearly 50 Democratic campaign ads included the word “fighter” or “fight”. Bob Chew, an independent Senate candidate in Colorado, has run an ad showing him throwing furniture across a room to express his rage over affordability.

A few candidates are taking a different tack. Josh Turek, a Democratic Senate candidate, says he will be “fighting for Iowa”, but also speaks of “the need for hope and positivity”. Mr Talarico talks about elevating service over selfishness. Last year Zohran Mamdani won New York’s mayoral race by pairing populist policies with relentless cheer. Americans are still cross—a growing share say they want change. Not every politician thinks the best way to offer it is with a furrowed brow. ■

ch078.xhtml

Pete Hegseth is conquering the Pentagon

His latest purge claims the army’s great reformer

Photograph: Getty Images

Sep 1st 2026 | Washington, DC

Ayear and a half ago few would have wagered that Pete Hegseth would consolidate power at the Pentagon. Narrowly confirmed by the Senate amid allegations of financial mismanagement in prior jobs and heavy drinking, which Mr Hegseth denied, the former Fox News host and voluble culture warrior appeared underqualified and no match for the generals and admirals he would direct. In office, the secretary of war has skirted scandal, be it mishandling secrets or forcing polygraph tests on his underlings. But far from seeing his power wane, Mr Hegseth has taken control of his department. A rolling purge of the armed forces has cleared out his rivals and those deemed insufficiently loyal.

Dan Driscoll, the army secretary, is his latest casualty. On August 31st Mr Driscoll tendered his resignation to the president after months of clashes with Mr Hegseth over the ouster of senior officers and the future of the force. Widely respected and viewed as a savvy reformer, Mr Driscoll’s departure raises yet more questions about the turmoil roiling the Pentagon and the viability of reforms meant to drive the army’s adaptation to the age of drones.

Mr Hegseth’s purges have been dramatic. Since taking office, he has ousted at least 24 flag or general officers. Many appear to have been dismissed for no reason beyond their race, sex or associations with officers Mr Hegseth disliked. A former national guardsman who rose to the rank of major, Mr Hegseth has long held a particular grudge against the army, an institution he claims “spit me out”. In April he sacked General Randy George, the most senior army officer and someone with whom Mr Driscoll had forged a close relationship. Chris Donahue, who headed US Army Europe and Africa, was forced into retirement two months later after Mr Hegseth downgraded his command. As of today the army has no Senate-confirmed chief of staff, secretary or commander of forces in Europe. Mr Driscoll’s deputy is also said to be on the cusp of leaving.

Mr Driscoll fought back as best as he could. “He prevented a number of officers from being fired or blocked,” says a congressional aide. Mr Hegseth’s recent removal of seven army officers from a list of 18 approved for promotion to two-star general may have been the final straw.

The departure comes at a pivotal time. Mr Driscoll, a confidant of J.D. Vance, the vice-president, is believed to have acted as a check on Mr Hegseth’s more hawkish instincts. He made a personal appeal to Donald Trump just days before his resignation, reflecting concerns in many parts of the Pentagon about the mounting strain on the armed forces from the war in Iran and Mr Hegseth’s meddling in the ranks.

Personnel churn has also undermined important modernisation efforts overseen by Mr Driscoll. He has been central to the Pentagon’s drone-procurement efforts (Mr Trump took to calling him “drone guy”). In November he visited Ukraine, forging a tech partnership with its drone specialists. The adulation Mr Driscoll attracted is thought to have upset Mr Hegseth, who is keen to dispense with initiatives associated with him and his allies. A new army battalion set up by General George to master the art of drone warfare, for example, is being phased out by General Christopher LaNeve, the acting chief of staff of the army, who is close to Mr Hegseth.

General LaNeve prefers to talk of a “back-to-basics” approach. While that still includes buying new tech, greater emphasis is being placed on MAGA-ish priorities. General LaNeve has overseen the drawing down of army units from Europe and placed greater focus on the Western hemisphere. Mr Driscoll had also fought with General LaNeve over his efforts to rewrite the army’s physical-fitness standards, a long-running fixation of Mr Hegseth’s.

The secretary is likely to solidify his hold further. Sean Parnell, the Pentagon’s spokesperson, is widely tipped as Mr Driscoll’s replacement. Like his boss, Mr Parnell seems enamoured of the culture wars. His lack of experience running large organisations could make confirmation difficult. But he could still serve in an acting capacity. With each passing day, the Pentagon looks more like Mr Hegseth. ■

ch079.xhtml

Where Canada can hurt America most

Its tariff retaliation is calibrated for maximum political pressure

In the war zone © Photograph: Getty Images

Sep 3rd 2026 | Washington, DC

THE THEORY behind Donald Trump’s trade war with Canada is simple: America holds all the cards. “WE DON’T NEED CANADA, THEY NEED US,” Mr Trump wrote on social media on August 24th. A week later Scott Bessent, America’s treasury secretary, was similarly dismissive of Canada’s retaliation. “I don’t think you can be in a tit-for-tat with someone who’s 13 times larger than you are.”

But that understates America’s exposure to one of its biggest commercial partners. The two countries traded more than $870bn in goods and services last year. Canada buys roughly 15% of all American goods exports and is the top export destination for 34 states. In carmaking, parts can criss-cross the border several times before a vehicle is finished. For America as a whole, the stakes of the trade war may look small. But for particular industries and states, they are anything but.

Canada is deliberately trying to turn those commercial ties into political pressure. Its response to America’s 50% tariffs on $20bn-worth of Canadian exports is a package of levies of 15-50% on $20bn-worth of American goods, due to take effect on September 8th. Steel and aluminium products are among those facing the heaviest tariffs, alongside cheese and agricultural equipment. Cars will continue to face existing Canadian counter-tariffs. “We are picking products that will target states,” said Mélanie Joly, Canada’s industry minister, on August 25th.

Canada’s strategy is clearest in the Midwest. Michigan, Ohio and Iowa all host crucial Senate races—and all will feel the pain. New 50% levies on steel and aluminium products will hit Michigan’s manufacturers, while the wider trade war weighs on the car industry. In Iowa, duties on agricultural equipment will hurt John Deere and other manufacturers. Ohio, meanwhile, sends nearly a third of its goods exports to Canada, making it particularly vulnerable.

Canada dropped seafood from its retaliatory list, sparing the lobster industry, but much of Maine’s trade remains closely intertwined with its northern neighbour. Lumber and paper products, among other things, routinely cross the border for processing. Susan Collins, the Republican senator facing a difficult re-election campaign, has called America’s tariffs “a mistake”. Farther south, 50% tariffs on furniture will hammer North Carolina, where another open Senate seat is in play.

Canada’s retaliation is not the only threat to Republicans. Mr Trump’s own tariffs will push up prices in America. The latest measures cover everything from toys to metals and other industrial inputs, raising the cost of both imported goods and American-made products. That is politically dangerous when affordability dominates voters’ concerns. Only 26% of Americans back Mr Trump’s tariffs on Canada.

Some Republicans are sounding the alarm. On August 24th Thom Tillis, the retiring Republican senator from North Carolina, warned that the trade spat was creating “tension” in border states. “The party that owns that tension”, he said, “is the party that probably suffers the consequences in November.” Others reject Mr Trump’s portrayal of Canada as an economic adversary. “This is not China,” said Ms Collins two days later. “It’s Canada, our best friend, a country with whom our economy is completely intertwined.”

The costs could yet grow. Jamieson Greer, America’s trade representative, has suggested outright bans on some Canadian goods. In Canada, other moves are being discussed, such as curbs on oil and critical-mineral exports. Mark Carney, the prime minister, says he is responding to an attack. Mr Bessent may be right that America has far more firepower. That does not mean it will escape injury. ■

ch080.xhtml

Woke 1 was about identity. Woke 2 is about class

That’s an improvement

Illustration: David Simonds

Sep 3rd 2026

The jury is still out on whether Mayor Zohran Mamdani, with his commitment to “the warmth of collectivism”, will summon a new era of plenty for struggling New Yorkers. But his young administration is already a gift to one constituency: conservative readers of the city’s tabloid, the New York Post, which is affirming their opinion of the left by feasting upon the privileged lifestyles of some of Mr Mamdani’s fellow travellers in the Democratic Socialists of America (DSA).

“Anti-rich DSA leader Gustavo Gordillo lives it up in posh $1.5m NYC pad paid for by millionaire parents,” the Post blared last month. Noting the DSA “railed against property ownership”, the Post revelled in the “lush landscaping” of the row house bought “by his mommy and daddy through a dummy corporation”. The Post later reported that Mr Gordillo, the co-chairman of the city chapter of the DSA, studied art and literature at Yale; that he was fired from an electricians’ union apprenticeship programme for failing to show up; and that his mother drives a silver Porsche.

Other leftist piñatas for the Post have included Tracy Rosenthal, who advocates abolishing rent and whose father is a successful music producer. (“Anti-rent radical and Mamdani pal who is ‘squatting’ in Brooklyn is Grammy nepo baby”.) The Post has branded Grace Ryan, “a 25-year-old chocolate heiress from Ohio”, as “one of the most shameless” members of the DSA for posting pictures of herself on social media “puffing a cigar on a boat in a $400 Ralph Lauren stars-and-stripes sweater”. To the _Post’_s unmistakable glee, Ms Ryan, who worked during the primary for a successful left-wing congressional candidate, defended herself on X by writing, “Like have y’all never heard of noblesse oblige?”

This is an old game. Since Karl Marx relied upon Friedrich Engels not just for his ideas about communism but for his textile fortune, lefties have fattened upon fat cats, and righties have made fun of them for it. From the start, socialism has also, for certain bourgeois faddists, radiated a subversive glamour. More rigorous socialists have seen this as the ideology’s Achilles’ heel. “The worst advertisement for Socialism is its adherents,” grumbled George Orwell in “The Road to Wigan Pier”. The typical socialist was either “a youthful snob-Bolshevik” or “a prim little man with a white-collar job” who had “vegetarian leanings” and, above all, “a social position which he has no intention of forfeiting”.

Yet Ms Ryan has a good point about noblesse oblige. She may never live like common people, but why should that bar her from trying to help them? Doesn’t it beat the alternative, a chocolate heiress in a $400 sweater who is indifferent to the plight of less fortunate Americans, if not out to screw them?

Here, amid the posturing, one arrives at something that may matter: an emerging distinction between the two periods of lefty activism becoming known as Woke 1 and Woke 2. Any such distinctions must be fuzzy and provisional, since few Democrats have had the decency to explicitly repudiate radical positions they once embraced. Some have shrugged off former excesses, as when Congresswoman Alexandria Ocasio-Cortez quoted another Democrat saying “Woke 1 was crazy!” and then laughed as though every American was in on the joke. Some have tried to redefine the slogans, insisting that “defund the police” did not mean defunding the police. Others have not retreated; time will tell whether they are the vanguard or the dead-enders. Still, some time between Donald Trump’s second inaugural address in January 2025 and Mr Mamdani’s first inaugural the next January, America’s left began to shift its emphasis, from race and gender to class.

Pick your causes or effects, whether the illiberal ideas that stampeded from the academy into social media; the rise of Mr Trump; the #MeToo movement; or the death of George Floyd under a policeman’s knee in 2020. These combined to concentrate the left on racism and gender bias. Identity became destiny. The poorest white person had privileges denied to the wealthiest black one. In theory the social-justice movement focused on policing; in practice it became obsessed with other preoccupations of America’s multiracial elite, such as preferential admissions to top schools. For prim little people with white-collar jobs, calling out advantages based on race became a way to prove enlightenment; so did disdaining the deplorables who voted for Mr Trump. Any true Marxist would have seen a classic attempt to split the working class, but Democratic leaders embraced this divisive politics.

Nothing to lose but more elections

Then came new effects and new causes. Kamala Harris lost in 2024 and Mr Trump demonstrated growing support among non-white working-class voters. Mr Trump cracked down on diversity, equity and inclusion programmes and the Supreme Court struck down race-conscious admissions. Democratic activists became outraged by injustice in Gaza instead of in America. Perhaps most important, Mr Mamdani won with an inclusive message about making life more affordable—as did moderate Democratic candidates for governor in Virginia and New Jersey.

For all the duelling between progressives and moderates, their agendas overlap when it comes to the precariat, if not the proletariat. From Georgia to Alaska, Democrats are all economic populists now, calling for higher wages, lower prices and more affordable health care. Victims and villains are judged not by the colour of their skin but by the content of their bank accounts. Billionaires are the bad guys, and in the capacious definition in use across the Democratic spectrum, anyone who works to pay their bills is working-class. This may not, in the end, revive the Democrats, let alone rescue struggling Americans. But if it does, the party will have at least some youthful snob-Bolsheviks to thank for helping remind it of the sorts of things it once stood for. ■

ch081.xhtml

America has laid claim to a fifth of Venezuela’s oil

The deal could give both countries reason to forestall a democratic transition

Photograph: AP

Aug 29th 2026 | Caracas

It was a typically Trumpian announcement. On August 28th the American president announced what he called “the biggest oil deal in world history”, an agreement he said would give America “majority control” of more than 65bn barrels of Venezuelan crude, around a fifth of the country’s vast proven reserves. He described Delcy Rodríguez, Venezuela’s unpopular leader, as the “Highly Respected Interim President of Venezuela”. Ms Rodríguez joined in the mutual obsequiousness, offering her “deepest gratitude” to Donald Trump for a deal she said would contribute to “our country’s economic growth”.

Mr Trump says the arrangement “more than doubles” American oil reserves. That is misleading. The United States, whose own reserves stand at around 46bn barrels, has not been “gifted” 65bn barrels of Venezuelan oil—as Mr Trump has claimed. Instead, via the Pentagon, it is taking a 35% stake in the parent company of a private oil producer, North American Blue Energy Partners (NABEP). The Venezuelan government has granted NABEP concessions over 14 oilfields, as well as three it already runs.

According to the White House, America will have the right to buy 20% of NABEP’s future production at cost, and will have first refusal for its remaining output. It will also have veto power over the appointment of any member of NABEP’s board of directors, the majority of whom must be US citizens. The Trump administration says the contract will be for 100 years and bring in more than $200bn in royalties and tax revenue for the Venezuelan state. The Venezuelan regime says the deal runs for just 25 years.

Mr Trump relishes the deal’s imperialist connotations. “To the victor go the spoils,” he said on September 1st. That referred to the raid carried out on January 3rd by American special forces to nab Venezuelan strongman Nicolás Maduro and his wife in Caracas, Venezuela’s capital, and fly them both to the United States, where they face drug-trafficking charges. Mr Trump suggests that once the “spoils” of that action—a large chunk of Venezuela’s oil—are in American hands, it will diminish the relevance of Middle Eastern oil, bring down fuel prices for American consumers and make Venezuela rich.

The long game

The reality is a lot more complicated. Significantly increasing production from Venezuela’s decrepit oil industry will take years. It is a rusting relic of what it once was after two decades of mismanagement and corruption under the administrations of the incompetent regime led by Mr Maduro and his predecessor, Hugo Chávez. “Venezuela will be completely irrelevant in terms of the global oil market price for the next year or so,” says Francisco Monaldi, a Venezuelan oil expert at the Baker Institute at Rice University in Houston.

Then there is the snag that Mr Trump is striking an agreement not with Venezuela, but rather its unelected leadership led by Ms Rodríguez, Mr Maduro’s vice-president, who has chosen to work with Mr Trump rather than suffer her ex-boss’s fate.

Chart: The Economist

On the streets of Caracas the main complaint is not that the deal is a colonialist grab by the United States, but that it props up a regime which Venezuelans voted out by a thumping majority in 2024. (That election, with the connivance of Mr Maduro’s cabinet, including Ms Rodríguez, was stolen). “This is all being done behind the people’s back, by people who are illegitimately in power,” says Xiomara, as she queues to fill her car with petrol.

That sense that Mr Trump is working with the latest incarnation of a rejected cabal is amplified by his choice of business partner. NABEP is controlled by Alejandro Betancourt, one of a group of Venezuelan businessmen who made fortunes on the back of lucrative contracts during the chaotic heyday of corruption under Mr Chávez. In Venezuela his name is forever tainted by an enormous contract granted to one of his previous companies by Mr Chávez in the early 2000s, to upgrade state electricity power plants. Venezuela’s political opposition called that deal over-billed and under-fulfilled, pointing to power cuts that plague the country to this day.

Hold your nose

“Betancourt is probably the most controversial person you could do this deal with,” says a person with connections to Venezuela’s oil industry. “He’s associated with the fact that we have one of the worst electricity services in the world.” Mr Betancourt, who has been investigated in Europe for money-laundering, has consistently denied all wrongdoing and has never been convicted of a crime. A US official calls him a man “who knows how to run this business” and pointed to his proven record in raising production in the three oilfields that his company already controls.

The American government insists that the deal with NABEP will involve no investment of public money. It therefore falls to the company itself to raise the $100bn it says is needed to quintuple its oil production to one million barrels a day. The fact that the US government has a stake in the firm, it is hoped, will give potential investors the security to commit.

The Trump administration is meanwhile pushing hard on many fronts the notion that Venezuela is ripe for investment. On September 2nd Mr Trump’s energy secretary, Chris Wright, was in Caracas for an announcement that Chevron, the only oil major to have remained in the country throughout the turmoil of the past quarter of a century, plans to invest $7bn in the next five years, in a deal separate from the NABEP one. But no other international oil company has made a similar commitment. Several have unhappy memories of Venezuela, losing billions when their previous joint-venture contracts were rewritten under the Chávez government in 2007.

Another source of their hesitation may be the question-mark still hanging over the durability of contracts. If a future democratic government comes to power, will it honour agreements forged under pressure with a government widely deemed illegitimate? “It will be a fiasco for all involved,” said Ricardo Hausmann, a former Venezuelan planning minister.

That in turn leads to a further troubling prospect. Both the Trump administration and the Rodríguez regime perhaps now share an incentive to stall democracy, given the risk it presents to their mutually cosy arrangement. “As the deal takes shape, the opposition’s role and influence could increasingly fade,” says Imdat Oner, a Turkish former diplomat who served in Caracas. Already, the Trump team has strongly discouraged the popular opposition leader, María Corina Machado, who left Venezuela to travel to Norway to collect the Nobel peace prize last year, from returning home. She has remained conspicuously silent about the deal.

The American official roundly rejects this, insisting that the Trump administration is rescuing Venezuela’s economy in advance of a future election. The plan is for the next government to inherit a country with a solid economic foundation, not one in crisis, so as to be better able to move beyond the catastrophe of the past 25 years. “A stable democracy, a stable republic, that can conduct elections on a repeated basis,” says the official.

Maybe. It was a Venezuelan energy minister who, in 1975, described oil as “the devil’s excrement”. Perhaps no deal involving the black stuff—let alone any nation-building—was ever going to be clean. ■

ch082.xhtml

The presidential candidate Brazil’s authorities tried to ban

Renan Santos, a fast-talking liberal with a rock band, models himself after Argentina’s Javier Milei

Rocking in the free world © Leonardo Carraro

Aug 31st 2026 | Rio de Janiero

Editor’s note (August 31st): Shortly after this piece was published, José Antonio Dias Toffoli, a Supreme Court justice who also sits on Brazil’s Electoral Court, suspended the candidacy of Renan Santos. Mr Toffoli said that the Santos campaign had failed to register its social-media handles correctly. The Santos campaign said it would appeal

RENAN SANTOS is cocksure. “I am the greatest orator of my generation in the world,” he claims in his small office, only half-jokingly. He has boasted of learning fluent Swedish in order to chat up girls on an annual pilgrimage to Europe that his friends have dubbed “the blondes tour”. He writes off his rivals as “infantile”, “pathetic” and “mediocre”. Yet for all his bragging, Mr Santos is no charlatan. The 42-year-old is running to become Brazil’s next president on a wonkish platform that sketches out a bold vision for the country.

Mr Santos’s rise has parallels with that of Javier Milei, Argentina’s firebrand president. Like Mr Milei, he is obsessed with free-market economics. In 2014 he founded a political group, the Free Brazil Movement (MBL by its Portuguese acronym), to advocate a minimal state in a country with a long tradition of big government. Like Mr Milei, Mr Santos mixes his dogmatism with rock music, fronting a band that extols the virtues of competition and low taxes. (Mr Santos claims to be the superior musician. His aide says both politicians’ bands are “equally awful”.)

And like Mr Milei, Mr Santos is a political outsider. Name recognition came after he organised mass protests against the government of Dilma Rousseff, a left-wing former president who was impeached by a hostile Congress in 2016. Only in 2025 did MBL register as a political party, under the name Missão (Mission). It has one congressman in office, Kim Kataguiri. As well as Mr Santos’s presidential run, the party is fielding 57 candidates for Congress in Brazil’s general elections in October.

Walking into the Missão offices reveals a project focused on cultural revolution as much as political power. The entrance is littered with copies of Revista Valete, the movement’s magazine, which has run features on Elon Musk and critiques of Judith Butler, a queer feminist philosopher. The mood is more startup than party headquarters. Programmers type code furiously behind large black screens, while down the corridor is a studio to shoot snappy videos for dissemination to the movement’s mostly male 4m online followers.

The focus on media is a response to “wokeism”, says Mr Santos. The rise of cancel culture “required responses that classic liberalism didn’t have to offer”, he explains. Many millennials like him, he says, went from supporting gay marriage and legalising drugs to “being labelled a fascist because you don’t want to use certain pronouns…or are expected to accept that a child can change their gender.”

Shouting for the underdog © Leonardo Carraro

But the cultural battle “is not our core identity”. Mr Santos distances MBL from Donald Trump’s MAGA movement. Economics remains its most important cause. The movement was spurred towards new thinking by “the failure of a childish form of economic liberalism” influenced by American libertarians, he says. “It is naive to think that if you remove the state from all activities and privatise everything, things will fix themselves. The state must be present in many areas, especially in a country like Brazil,” with gaping inequalities. The party now believes that companies such as Petrobras, Brazil’s state-owned oil firm, should not be privatised, but that the government’s influence should be reduced. Mr Kataguiri, who oversees the party’s economic proposals, calls this “pragmatic, non-dogmatic liberalism”.

Mr Santos wants to turn Brazil into a great power within 30 years through deep structural reforms. The transformation is sketched out in the party’s “Yellow Book”, a 350-page manifesto which applies lessons for Brazil from successful public policies around the world. Mr Kataguiri travelled to Denmark to study its patent laws and to Japan to learn about its system of evaluating local governments. Germany’s apprenticeship system, China’s infrastructure spending and the United States’ light-touch approach to AI regulation all feature. “We definitely don’t want the EU model, which is more obsessed with controlling and regulating than with developing its own AI,” says Mr Kataguiri.

The result is a policy programme packed with new ideas for Brazil. Missão proposes digitising patient health-care records and using AI to improve triage and reduce queues at public hospitals. A new law would create performance targets for local politicians on education, health care and access to sewerage. Officials who hit targets would be eligible for more money from the federal budget; persistent underperformers would be threatened with a temporary prohibition from holding office. AI would be used to recommend cuts to public jobs. “There is no limit to how AI can be used to improve administrative efficiency,” gushes Mr Kataguiri.

Tackling the big one

Mr Santos is frank about the need for spending cuts. Over 90% of Brazil’s federal budget is tied up in mandatory spending, particularly pensions. The share is growing relentlessly as many pensions and several welfare benefits are indexed to the minimum wage, which rises automatically with inflation and GDP growth. By 2029 discretionary spending is expected to comprise an unworkable 0.1% of the budget. Missão proposes de-indexing pensions and benefits and tying spending on education and health care to inflation rather than revenues, which are growing fast. Cuts to lavish civil-servant salaries would follow, and a pension reform. All this, hopes Mr Kataguiri, would make room for growth-boosting infrastructure investments, which he wants to double to 4% of GDP.

When it comes to security, populism trumps liberalism. One of Mr Santos’s slogans is “Prendeu, Matou” (“Catch, Kill”), referring to how he believes police should treat criminals. He frequently praises Nayib Bukele, El Salvador’s strongman, who has reduced crime by locking up 2% of the country’s adult population and trampling over democracy. (Mr Santos claims to be “Bukele in style, Milei in substance”.) He promises to call states of emergency and allow a “bloodbath” against gangs. The Yellow Book extols “the doctrine of the enemy”, a controversial legal theory which argues that some criminals do not deserve civil rights. In practice it could be easily abused by an executive bent on repressing its political enemies.

Mr Santos has been polling in the single digits, but his willingness to propose bold ideas is winning him prominent backers on the Brazilian right. Many are disheartened by the candidacy of Flávio Bolsonaro, a son of Jair Bolsonaro, a populist former president who is serving a 27-year sentence for attempting a coup. “Flávio is extremely weak,” says one Santos supporter. “Renan is well-read, informed and can speak about lots of different subjects. Flávio doesn’t know anything.” Mr Santos attacks the Bolsonaro family for being intellectually mediocre and corrupt.

Mr Santos will struggle to win over women. He has praised manosphere bloggers. One of his closest allies, Arthur do Val, with whom he travelled to Ukraine in 2022, has said that female Ukrainian refugees were “easy because they’re poor”. On that trip, Mr Val boasted that Mr Santos had given him tips on how to pick up women. “You’ve got to go to ordinary towns…these poorer towns are the best.”

On August 31st José Antonio Dias Toffoli, a Supreme Court justice who also sits on Brazil’s Electoral Court, suspended Mr Santos’s candidacy on the grounds that he had failed to register his party’s social-media handles correctly. A day later, after an uproar, Mr Toffoli reversed course. But the attempt, which many experts considered extreme, has boosted Mr Santos’s anti-establishment cred. Even if he does not win this time, he has given fed up right-wing voters an alternative to bolsonarismo. ■

ch083.xhtml

A prawn superpower rises

Ecuador’s farmers have embraced technology to great effect

Grey gold © Reuters

Sep 2nd 2026 | Guayaquil

The production of prawns in Ecuador is lean manufacturing incarnate. Vast rectangular pools tessellate the Guayas delta, centre of the aquaculture belt. Its brackish waters hide billions of translucent bodies. Larvae are cultivated in greenhouse-style hatcheries, fattened in open-air ponds, hoovered into lorries, then taken to packaging plants where they are frozen. Inefficiencies are constantly hunted down. Containers filled with bricks of prawns known as “HOSO” (head-on, shell-on) cross the oceans, mostly east to west across the Pacific. “It is like a factory,” says Wolfgang Harten of Grupo Almar, the country’s second-largest producer.

Chart: The Economist

That factory is running flat out. Shipments have quadrupled in the past decade. In 2020 Ecuador surpassed India as the world’s top exporter (see chart). Last year it sold more than $8bn-worth of prawns, more than India, Vietnam and Indonesia combined. Once an oil nation, Ecuador has become a pesco-state. Prawn production provides employment for about 1.6% of the population and furnishes 6% of GDP, a larger share than carmaking contributes in Germany. The sector has developed the country’s most advanced technology. The lorries that move prawns are escorted by armed guards. For coastal communities, a job at a prawn-processing mill is a ticket to the middle class.

Business wasn’t always this good. When Ecuador began cultivating prawns in the 1960s it was a cottage industry. Penaeus (Litopenaeus) vannamei, better known as king prawn or white-leg shrimp, is native to Latin America. Farmers collected wild larvae from nearby mangroves. Few bothered with selective breeding. Profits were high enough to drive expansion, but at great cost. By 2000 deforestation had halved the country’s mangroves. A global outbreak of white-spot disease bankrupted many farmers, killed 95% of prawns in infected ponds and cut output by 70% over two years.

Ecuador’s Asian rivals responded to the disease by intensifying production. Ponds were crammed with densities as high as 300 prawns per square metre. Larvae were isolated in labs to ensure they had never contracted white spot. Antibiotics became routine. Ecuador took the opposite path. Instead of trying to eliminate disease, producers adapted to it. Ponds were thinly stocked, with densities a tenth of Asia’s. Larvae were bred from the survivors of white spot, which had acquired resistance to it. Antibiotics were phased out.

The country improved its technology. Since the 2010s waves of innovation have put its farms at the frontier of marine tech. Many farmers in Asia still row out in a canoe twice a day to toss in a bag of feed. In Ecuador they own automatic feeders that drop pellets every three minutes. Underwater microphones record the click-click of chewing crustaceans to tell when they are full. Depending on consumption speed, ingredients are tailored for the next cycle, says Andrés Rivadulla of BioMar, a feed manufacturer. Aerators pump oxygen through each pond. Mechanical filtration reduces water pollution.

The results have been phenomenal. The combination of healthier prawns and better tech means that, although productivity is below that in places like Thailand and Indonesia, survival rates are far higher. Around 95% of prawns are harvested alive in Ecuador, compared with barely 80% in Asia, says Gorjan Nikolik of Rabobank, a Dutch lender. Minimising disease outbreaks lowers costs and stabilises yields. Cultivating a native species helps. Grupo Almar manages six annual cycles of 50 days, nearly twice the pace of competitors in India and Vietnam. This has helped consolidate prawn firms and integrate family farmers into well-capitalised corporations.

Geopolitics have also helped the South Americans. Ecuador is already the biggest supplier of prawns to the European Union and China. Yet it had long struggled to gain ground in North America, where India’s output dominated. Then along came Donald Trump. Ostensibly angry about India buying Russian oil, last year he hit its goods with a 50% tariff. Facing a combined duty of nearly 60% on Indian prawns versus 10% on the Ecuadorians, importers in the United States switched en masse. Americans now consume more prawns from Ecuador than from anywhere else; Ecuador kept its lead even after India’s duties were relaxed earlier this year.

Don’t overdo it

In fact the main problem is over-production. Since recovering from the white-spot epidemic, farmers have boosted output by 15% per year. This has swamped markets, dragging prices to record lows. A kilogram of prawns cost $2.40 at the farm gate in Ecuador in May, compared with $2.99 in India and $5.32 in China. To shield its own producers, China sporadically restricts imports, claiming phytosanitary concerns. Brazil’s import duties are prohibitive. Mexico complains of smuggling. Then there is climate change, which actually increases oversupply. By raising temperatures on the Pacific coast, the El Niño weather phenomenon harms crops but stimulates prawns, which love heat.

So the priority is to boost demand. The Balkans and the Persian Gulf are seen as promising markets for HOSO. Companies are starting to make higher-value products for American consumers, such as peeled and breaded prawns. Fast-food shops are promoting prawn burgers while marketers extol the healthiness of prawn-based ceviche, tempura and risotto. “The best way to start your day is with a prawn omelette,” says José Antonio Camposano of the National Chamber of Aquaculture, an Ecuadorian trade-group. It may be working. Demand is up. But Ecuador’s prawn kings are hungry for more. ■

ch084.xhtml

What will happen to Uganda when Yoweri Museveni goes?

Nothing good, if recent months are anything to go by

Illustration: Tarini Sharma

Sep 3rd 2026 | Kampala

On two sets of posters plastered across Kampala, Uganda’s capital, two men loom large. One is a paunchy figure with a wide-brimmed hat and a jaundiced stare. The other is younger and brasher-looking in military fatigues. The first is Yoweri Museveni, the 81-year-old president, who seized power after a civil war 40 years ago and is entering his dotage. The second is Muhoozi Kainerugaba, Mr Museveni’s capricious son.

General Kainerugaba serves as Uganda’s army chief, but he covets his father’s crown and increasingly behaves as though he were already in charge. Older Ugandans say his belligerence and brutality are reminiscent of Idi Amin, an infamous military dictator of the 1970s. His ambitions are causing ructions within the army and the ruling party, and even his own family, raising fears of a power struggle within the regime when Mr Museveni eventually goes. “Ours has been a history of extremes: either you have anarchy, or real totalitarianism,” says Norbert Mao, Uganda’s minister of justice. He is not alone in worrying that those days could soon return.

Mr Museveni put General Kainerugaba in charge of the armed forces in 2024. Yet even if the president has indeed been grooming his son as a successor, the two men are quite different. In person, Mr Museveni is a charismatic storyteller; General Kainerugaba tends to be taciturn. The father is fond of long speeches; his son is notorious for undignified outbursts on X. “These days I’m as sober as a cat,” he posted in the early hours of September 1st. “Women need to stand at least ten metres from me…for their safety.”

Outsiders sometimes see General Kainerugaba, who is 52, as a sort of comically conceited nepo-baby. Yet few Ugandans are laughing. During a crackdown after Mr Museveni’s re-election in a flawed vote in January, the general threatened to kill and castrate Bobi Wine, the main opposition candidate. Mr Wine went into hiding and is now in exile. Kizza Besigye, another prominent opposition figure, was abducted from Nairobi, Kenya’s capital, in 2024 and has been in prison ever since. He is standing trial on treason charges; General Kainerugaba has vowed to hang him. During a court hearing in July, Mr Besigye collapsed mid-speech and was hospitalised.

The general’s depredations have intensified in recent months. In July his soldiers abducted Erias Lukwago, a former mayor of Kampala and one of Mr Besigye’s lawyers. The army chief posted pictures of a blindfolded Mr Lukwago on X, boasting that he was in his “basement”. Mr Lukwago was later accused of treason and remains incarcerated. Between mid-June and mid-July alone, security forces unlawfully detained at least four other dissidents, according to Human Rights Watch. Several are believed to have been tortured.

Repression in Mr Museveni’s Uganda is nothing new. But there is a widespread sense that the regime is degenerating into something darker and more despotic. The state once sought to cover its abuses with a veneer of constitutional legitimacy. “Now we have a rabid, unapologetic, in-your-face dictatorship,” says Eron Kiiza, a lawyer who was arrested and tortured last year and is also now in exile.

Such views are not limited to the opposition. “The gap between what we promise in our laws and what is really happening is widening,” admits Mr Mao, the justice minister. He says the army has violated the constitutional rights of Mr Besigye by denying him medical treatment when he is evidently ill. The minister has also received credible reports of torture by the security forces. But General Kainerugaba “doesn’t care what the law says”, Mr Mao observes. “I don’t know whether we have reached the level of Idi Amin’s time, where bodies are found floating on the Nile,” says another government insider. “I think not yet. But these are grave issues of concern.”

Can anything stop the slide? Mr Museveni has never explicitly said he wants his son to succeed him. In the past he has reined in some of his wilder escapades. But the leader has grown frail of late. He increasingly seems unwilling or unable to restrain his son.

That appears to have emboldened the general to try to crush his perceived enemies and rivals. In May he had Anita Among, a powerful former speaker of parliament once considered untouchable, put under house arrest (she has since been released). Several senior officers have been purged. On August 25th the general’s men kidnapped a close associate of Odrek Rwabwogo, Mr Museveni’s son-in-law, who is believed to have presidential aspirations. “This is my last warning to that fool Odrek,” fulminated the general on X, threatening to arrest Mr Rwabwogo and to have the associate either beaten to death or executed by firing squad.

The one man who could perhaps restore some sanity to Mr Museveni’s court is Salim Saleh, the president’s reticent younger brother. A retired general, Mr Saleh is a powerful figure behind the scenes. He is understood to be disconcerted by some of his nephew’s behaviour, and is considered the only person aside from Mr Museveni whom General Kainerugaba dare not defy.

If a dynastic succession cannot be avoided, much of Uganda’s elite would prefer Mr Saleh as president. He is seen as rational and comparatively conciliatory on ethnicity. General Kainerugaba, by contrast, is widely reviled as an ethnic chauvinist. He calls his own people, the Bahima, “a special breed”, disparaging other groups with crude racial slurs. Many Ugandans are also suspicious of his close ties with neighbouring Rwanda. The general likes to say he has two uncles: Mr Saleh and Paul Kagame, Rwanda’s strongman president. “Which uncle will eventually have more influence on his student?” asks the government insider.

It is unclear what Mr Saleh wants. Some think he would prefer to remain a kingmaker. Others reckon that in the event of a crisis within the regime, he would intervene and, if necessary, take power himself. “The Raúl Castro scenario is not far-fetched,” says a well-connected observer.

How General Kainerugaba would react is anyone’s guess. Many suspect that the bulk of Uganda’s army, if push came to shove, would side with Mr Saleh. Yet Uncle Kagame, across the border, may have his own ideas about how things should play out. Mr Saleh and the general are said to be close, making it difficult to imagine they would come to blows. But when so much power is concentrated in so few hands, a single family dispute could be enough to destabilise a country. ■

ch085.xhtml

Ethiopia’s undeclared drone war

A spate of strikes on Tigray pushes the country into dangerous territory

You will hear of wars and rumours of wars © Liu Fangqiang/Xinhua/Eyevine

Sep 3rd 2026 | Nairobi

IN TIGRAY, ETHIOPIA’s northernmost province, August marks the start of Ashenda, an Orthodox Christian festival. It is a time of family reunions, street parades and concerts. But this year public parties were cancelled and private ones were muted.

On August 20th the Ethiopian government attacked Mekelle, Tigray’s capital, with fighter jets and drones. The strikes reportedly injured 13 people, including children. From August 24th to August 27th, Tigray was bombed at least once a day. One drone strike in the south on August 26th destroyed a school. “They are trying to terrorise and humiliate the public,” says a senior official in the Tigray People’s Liberation Front (TPLF), Tigray’s ruling party. “The only way out is to fight back.”

Map: The Economist

The escalation has been a long time in the making. Between 2020 and 2022 the Ethiopian government and the TPLF fought a brutal civil war, which is thought to have killed hundreds of thousands of people. The peace deal that ended the fighting failed to resolve the core of the dispute between Abiy Ahmed, Ethiopia’s prime minister, and the TPLF’s leaders, who had dominated Ethiopian politics for decades until Mr Abiy elbowed them aside in 2018. In recent months the Ethiopian army has engaged in several skirmishes with forces loyal to the TPLF. To many observers, a return to civil war has long seemed almost inevitable.

Yet the recent attacks are also puzzling. Unlike during the previous war, the Ethiopian government has not cut off telecommunications in Tigray. Commercial flights between Addis Ababa, Ethiopia’s capital, and Mekelle continue. In 2020 Mr Abiy declared war on the TPLF on national television before sending the tanks in. This time his government has not even commented publicly on the drone strikes. Instead, months of unresolved tensions have quietly given way to what is in effect an undeclared blitz. Tigray is now a grey zone: no longer at peace, if not quite at war.

Part of the reason for the ambiguity may be that Mr Abiy is more cautious this time round. His bellicose rhetoric in 2020, as well as a slew of atrocities committed by Ethiopian troops, tarnished his one-time reputation as a peacemaker. Crucial economic support from the IMF and the World Bank could be imperilled if he were blamed for starting another war. Fortunately for Mr Abiy, the TPLF has auditioned energetically for the role of perceived aggressor. It has been forcibly conscripting young Tigrayans for months, threatening to “bury” Mr Abiy’s regime.

Mr Abiy may also be reluctant to rush in because regional alliances have shifted. In the previous war, militias from the Amhara region, to the south of Tigray, and troops from Eritrea (to its north) fought alongside Mr Abiy against the TPLF. Both have since formed an improbable alliance with their erstwhile foe. Eritrea, in particular, has emerged as Mr Abiy’s chief antagonist. Although Isaias Afwerki, its veteran dictator, loathes the TPLF, he is now more worried about the Ethiopian prime minister’s desire to annex part of his coastline. “The Eritreans are not stupid,” says the TPLF official. “They see the regime in Addis as an existential threat.”

Neither side seems interested in peace talks. The TPLF says the deal that ended the last war is “dead”. Ethiopian state media say the TPLF is “defunct”. On August 24th the TPLF’s chairman half-heartedly met Olusegun Obasanjo, a former president of Nigeria and one of the mediators of the moribund accord. The TPLF said the discussions “produced neither new proposals nor evidence of tangible progress”.

What next? Loth to launch a ground war against multiple foes on multiple fronts, Mr Abiy seems likely to focus on Tigray’s skies. The Ethiopian army has near-total air superiority over the TPLF. It has relied heavily on aerial bombardment in its war against rebels in Amhara, where many civilians have died. Last year a drone strike in East Gojam killed 16 people, including an eight-year-old boy. Such incidents have become so common in Amhara that they barely attract media coverage.

The TPLF, for its part, has vowed to reconquer Western Tigray, a swathe of fertile land that was ethnically cleansed by militias allied with Mr Abiy during the previous war. Yet many doubt whether it can, having been weakened by years of infighting. Several political leaders and military commanders have defected to Mr Abiy since the start of 2025. This has doubtless given his government a wealth of intelligence on the TPLF’s military infrastructure and planning. Some observers reckon the relentless drone strikes have left the group in disarray. “They are too weak to respond,” says an analyst in Addis Ababa.

Much will depend on how much support the TPLF gets from Eritrea—an uncertain prospect, given its long history of enmity with Mr Isaias. It may also be hoping for cross-border support from the Sudanese Armed Forces (SAF), Sudan’s national army, which resents Mr Abiy for backing its rival, the Rapid Support Forces, in Sudan’s civil war.

On August 1st the TPLF and the Ethiopian army clashed in Shererina, a pocket of Western Tigray on the border with Sudan. A think-tank with close ties to the Ethiopian government later accused the SAF of aiding the TPLF. Last week the SAF said it had shot down three hostile drones launched from Ethiopian territory. The TPLF’s leaders insist that Eritrea and the SAF are committed to removing Mr Abiy from power. The grey-zone crisis in Tigray may yet paint the region red. ■

ch086.xhtml

Iran’s proxies are going their own way

After three years of war, the axis of resistance is no longer an axis

Photograph: Getty Images

Sep 3rd 2026 | DUBAI

For years Iranian officials spoke of an “axis of resistance”, allies from Beirut to Sana’a who would stand together against common foes (see map). Yet when Iran was attacked twice in the past 15 months, each made a different decision about whether to join the fray. The axis may not be dead—but the brand has outlived the business.

Map: The Economist

The strategy of backing Arab militias dates to the 1980s, when Iran sent its Islamic Revolutionary Guard Corps (IRGC) to train Hizbullah fighters in Lebanon. It saw such groups as a substitute for its weak conventional army: they would serve as forward defence, projecting power while keeping conflict away from Iran.

The defence clearly failed. The Hamas-led massacre on October 7th 2023 started a cascade of wars that ended with Iran itself being attacked. Hizbullah and Hamas are badly weakened. Bashar al-Assad’s regime in Syria has gone. The Iraqi government has ordered militias to disarm by October.

Some regime foes hope that Iran will have to curtail its support, perhaps to facilitate a deal with America, and trigger the collapse of the axis. That is unlikely. The militias are wounded but alive, and the strategic rationale for supporting them is stronger than ever. Yet the axis may also struggle to rebuild itself as a consortium: the parent company in Tehran is bankrupt, and many of its subsidiaries are poorly run.

A likelier outcome is that its constituent parts survive but diverge. Some will become Iranian wards, held tightly but used sparingly. A few will graduate from the axis. Others will emerge as strays: clients that might have somewhere else to go. Counter-intuitively, Iran will exercise the tightest control over groups that have suffered the greatest blows to their capabilities.

Take Hizbullah, once Iran’s most feared proxy. Until 2023 it had an arsenal of more than 100,000 rockets and missiles aimed at Israel and a well-trained commando force that had honed its skills fighting to save Mr Assad’s regime. Hassan Nasrallah, its charismatic leader assassinated by Israel in 2024, saw himself as a peer rather than a proxy. He was a true believer steeped in Iran’s revolutionary ideology, but tried to preserve some autonomy.

No longer. Naim Qassem, Nasrallah’s successor, is an unloved middle-manager. Israeli strikes have killed many of the group’s experienced commanders. The Lebanese government insists that Hizbullah must hand over its diminished arsenal. With so many commanders dead, the IRGC sent its own men to Lebanon to exercise direct control over the group, leaving it little choice but to start another war with Israel in March. At the same time, its domestic popularity has collapsed. Iran will preserve Hizbullah less for what it can do than for what its loss would signal.

At the other extreme are the Houthis, the most independent-minded of Iran’s allies. They started as a local insurgency in northern Yemen in the 1990s. It would be decades before Iran offered significant financial and military aid. When it came, it vastly improved their capabilities. Once derided as barefoot gunmen, Houthi fighters now wield anti-ship missiles to harass vessels in the Red Sea and air defences to shoot down American Reaper drones.

Photograph: Getty Images

Yet the Houthis continue to guard their autonomy. Their involvement in the post-October 7th wars has waxed and waned, based more on their own interests than Tehran’s. They are increasingly able to produce their own weapons and are forging their own alliances, for example with al-Shabab, the al-Qaeda affiliate that controls much of Somalia. Houthi delegations have visited Russia and held talks with China. Iran and the Houthis will continue to find common interest: both loathe America, Israel and Saudi Arabia. But that is not the same as taking orders.

In between are the strays. Many Shia militias in Iraq are a by-product of the American-led invasion in 2003, after which Iran poured weapons and money into the country to bleed its American foes. Two decades later their priorities have changed. Iran remains an ally, but their main patron is the Iraqi state, which has more than 200,000 militiamen on its payroll.

Some militias did launch attacks on American bases earlier this year, but they worry that open conflict with the Iraqi government or its allies would jeopardise their position. The IRGC is said to have sent advisers to Iraq to establish small cells tasked with attacking Gulf states, a sign that the bigger militias were reluctant to do so.

Hamas was always an awkward fit in the axis, a Sunni group in a largely Shia bloc. It fell out with Iran over the latter’s support for the Assad regime in Syria. Khalil al-Hayya, its newly appointed leader, is considered a loyal Iranian ally, but other officials would prefer a pivot towards Sunni powers, namely Turkey and the Gulf.

In March Hamas urged Iran to halt its attacks on Gulf states. Iran will not abandon it—support for the Palestinians is central to the regime’s identity—but its support may be more rhetorical than material.

Even if Iran wanted to rebuild all its allies, its resources are limited. The war has caused hundreds of billions of dollars’ worth of damage. America is blockading its ports and tightening sanctions on its economy. The fall of Mr Assad’s regime also severed the land route to Lebanon. Iran will need to prioritise.

In some ways, the fate of the axis may mirror that of Soviet client states in the 1980s, when Moscow could no longer afford to offer them generous support. This is an imperfect analogy, of course. Moscow’s allies had a rival superpower to turn to, while Tehran’s face a messier board: some have a range of possible partners; others, none at all. Still, the Middle East’s most ambitious conglomerate may be heading for a restructuring. ■

ch087.xhtml

Binyamin Netanyahu sows discord ahead of Israel’s election

The campaign is entering a new phase

Photograph: Reuters

Aug 31st 2026 | JERUSALEM

TROJAN HORSES seem to be everywhere this summer. Binyamin Netanyahu, Israel’s prime minister, has featured one in a campaign video for Israel’s parliamentary election on October 27th. In the ad, an AI-generated Gadi Eisenkot, a former general now leading the polls, rides the horse. Lurking inside are Mansour Abbas and Yair Golan, two other opposition leaders.

Mr Netanyahu’s message is crude—and clear: Mr Eisenkot can win only by joining forces with Mr Abbas’s Arab-Israeli Ra’am party and Mr Golan’s Zionist-left Democrats. Yashar, Mr Eisenkot’s centre-right party, has overtaken Mr Netanyahu’s Likud in The Economist’s poll tracker. But he will need to woo a range of partners to form a government—as will Mr Netanyahu.

Mr Eisenkot’s growing popularity among centre-right voters, Likud’s constituency, is unsettling Mr Netanyahu’s party. Hence it is trying to taint Mr Eisenkot by association. Smearing the Arab parties, which represent over 20% of Israel’s population, is an old tactic. Mr Netanyahu has repeatedly labelled them “terror supporters” and accused Ra’am of being an offshoot of the Muslim Brotherhood (though he has not been above trying to elicit Mr Abbas’s support when desperate).

Efforts to do the same to a Jewish and Zionist party are new. A decade ago, when Mr Golan was deputy chief of staff of the Israel Defence Forces, he likened what was happening in Israel to “horrifying processes that occurred in Europe” before the Holocaust. More recently, during the war in Gaza, he said that “a sane state does not wage war against civilians, does not kill babies as a hobby, and does not set goals for itself like the expulsion of a population.”

Mr Golan’s opponents are now using these quotes to depict his party as radical and non-Zionist, and therefore unfit for government. But so far the polls are not moving in Mr Netanyahu’s favour, and none of the main opposition parties has ruled out governing with the Democrats.

When it comes to including an Arab party in the next government, however, Likud has managed to split the opposition. Mr Golan is the only one openly in favour. All his potential partners are either adamantly opposed to working with the Arab parties or have been evasive. In public Mr Eisenkot has said he will consider governing with any party that accepts a vague set of principles. But he would rather lead a minority government, however unstable, than include Israel’s Arab representatives.

Though his alliance of hard-right and religious parties may not be capable of winning a majority, Mr Netanyahu is already dictating the terms by which his rivals will try to form a coalition. His campaign is pushing Mr Eisenkot towards harsher nationalist positions; the latter recently ruled out the possibility of accepting a Palestinian state if he were to govern. He may struggle to form anything better than a short-lived government bound to policies not unlike those of the outgoing one.

Mr Abbas, one of the canniest politicians in Israel, is the only Arab-Israeli leader to have ever joined a ruling coalition. In 2021 he supported a government led by Naftali Bennett, a right-wing politician who is running again this year. Mr Abbas claims the other Arab-Israeli parties, which are running together on a joint list, are refusing to make the necessary compromises to be in government and effect change.

To boost his chances of joining a government again, Ra’am will field its first Jewish candidate this year. Yoav Segalovitz is an unlikely pick for a party dominated by conservative Islamists. A secular centrist, Mr Segalovitz’s main attraction is his former career as a police investigator. One of Mr Abbas’s priorities is to get the police, currently under Itamar Ben-Gvir, the leader of the ultranationalist Jewish Power party, to take seriously the bloody wave of gang violence afflicting Israeli Arabs.

But he also wants to make Ra’am more palatable to Jewish parties. Mr Segalovich could help. Like Mr Netanyahu, Mr Abbas is not only trying to win votes, but to change the terms of coalition-building. Rather than languish as a hidden passenger within the Trojan horse, Mr Abbas wants to ride it. ■

ch088.xhtml

How the Alternative for Germany would govern

If it wins its first state, ideology in schools and broadcasting is just the start

Photograph: HANNIBAL HANSCHKE/EPA/Shutterstock

Aug 31st 2026 | MAGDEBURG AND ZERBST

IT IS A horribly hot afternoon in the market square of Zerbst, a small town in the east German state of Saxony-Anhalt, but Ulrich Siegmund’s grin barely slips. For three hours Mr Siegmund (pictured), lead candidate for the populist-right Alternative for Germany (AfD) in Saxony-Anhalt’s election on September 6th, signs t-shirts and takes selfies with fans who have endured long waits under the blistering sun. There is no trace of the resentment that once typified AfD rallies. “This is not a party, it’s a movement!” beams Phillipp-Anders Rau, the local AfD hopeful.

Chart: The Economist

Mr Siegmund, whose slogan is the sunny “Everything is possible!”, hopes to become the first right-wing nationalist leader of a German state since the second world war. The AfD is set for a big win over the ruling Christian Democrats (CDU) on Sunday (see chart). Depending how many other parties cross the 5% threshold, it could secure a majority alone. If it falls short Mr Siegmund has other routes to power, perhaps via CDU defections—although he insists to your correspondent that he has no interest in that.

Supporters in the crowd dismiss allegations that the party has dark designs. “The AfD is not radical, it just stands for freedom of opinion,” says Heike Schmidt, a local. Yet its plans for Saxony-Anhalt, laid out in a 258-page governing programme, are among the most sweeping ever produced by a party with a credible chance of taking office. They amount to a vision for a new sort of state, one party officials see as a blueprint for the rest of Germany. But they are also uncosted and in places unserious.

Some proposals, such as abolishing the right to asylum or restoring relations with Russia, are not in the gift of state governments. But plenty are. Start with education. The AfD wants to remove disabled children and the children of refugees from regular classrooms, which experts say violates European and international law. It wants more Russian-language teaching, and to restore exchanges with schools in Russia. It aims to remove the Schulpflicht, or obligation to attend school, so that distrustful parents can educate their children at home. The AfD would rewrite curricula to excise wokery and celebrate historical German achievements. Hans-Thomas Tillschneider, the state party’s chief ideologue and principal author of its manifesto, argues that Germany’s culture of condemning the Nazi past has induced a national “identity disorder”. “More Bismarck, less Hitler” is the remedy.

The AfD detests Germany’s universities in their current form. Mr Tillschneider has called academic humanities the “legitimising whores” of social change. The party wants to restructure university governance, to withdraw from European standards for granting degrees and to establish a chair of demographic studies and an institute for “critical Islamic studies”. It will struggle to achieve much of this: academic freedom is constitutionally protected. But since universities are largely funded by the states, an aggressive government could apply financial thumbscrews.

Mr Siegmund has promised to replace the public broadcaster, long a bugbear of Germany’s populist right, with a “neutral” alternative. The AfD’s programme takes aim at Saxony-Anhalt’s slogan #moderndenken (“think modern”) and its celebration of the 20th-century Bauhaus architectural movement, based in Dessau. The party sees Bauhaus as just the sort of deracinated culture it wishes to eradicate. Its alternative approach, #deutschdenken, emphasises the achievements of Martin Luther, Friedrich Nietzsche and the Holy Roman Empire, and proposes “patriotic tourism concepts”.

Yet getting all this done assumes a degree of competence that many, even inside the AfD, suspect is lacking. German states require hundreds of officials and staff, and bar a couple of towns and districts the AfD has no governing experience. Two of Mr Siegmund’s closest aides once belonged to a neo-Nazi youth group; unpromising material for running a state. A “Black-Red-Gold” academy set up last year is training party cadres in administration, and the AfD has sought advice from Austria’s populist-right Freedom Party, which has joined governments at both the state and national level. But this does not seem equal to the task.

Fiscal constraints will also hold the AfD back, notes Petra Dobner of the University of Halle. Saxony-Anhalt, a poor state of 2.1m people, owes around €25bn ($29bn) and relies on huge transfers from the rest of Germany. The AfD’s programme, which includes lavish promises for child care and school meals, would leave an annual financing gap of €2.2bn, according to the Halle Institute for Economic Research. Perhaps in acknowledgment, the AfD’s recent 100-day governing programme focuses on symbolic matters like putting up German flags and scrapping gender-neutral language.

Yet organisations in the party’s crosshairs are not complacent. Universities, judges and NGOs have for months been drafting strategies to prepare for an AfD government. Earlier this year Saxony-Anhalt lawmakers changed their rules for electing judges and the parliament’s president. A big fear is that the AfD could push through endless anti-constitutional measures in the knowledge that courts would take years to overrule them, sapping adversaries’ morale.

The prospect of an AfD government is concentrating minds elsewhere in Germany. Noting the party’s cosiness with Russia, interior ministers in the 15 other states have been mulling how to exclude it from some intra-ministerial discussions. Georg Maier, the interior minister of Thuringia, favours erecting “Chinese walls” to ensure crucial national-security information is not shared with the party’s ministers.

The AfD also fears opposition from the “deep state” within Saxony-Anhalt. Mr Siegmund says he would seek to replace up to 200 state employees, and has threatened “administrative measures” against others who stand in his way. The party may make a handful of political appointments, and can fire some officials. It will surely dismiss the head of the state’s intelligence service, who in 2023 decreed the Saxony-Anhalt AfD “right-wing extremist”. Most civil servants cannot simply be fired. But Friedrich Zillessen, a legal scholar who works on judicial resilience in the state, notes that an AfD government could find other means: attrition as civil servants retire, promoting friendly figures or creating a culture of intimidation.

Saxony-Anhalt is hardly a model of health. It has the highest median age among Germany’s states and some of its worst poverty. The AfD promises to replace “culturally alien” professionals by seducing Germans who have left the state to return. But its policies seem likely to have the opposite effect. Talk is rife of workers who might leave if the AfD takes office. Outsiders, especially foreigners, who might otherwise have been tempted to come will surely think twice. The AfD has even vowed to reject its share of a national infrastructure fund established last year by Friedrich Merz, the chancellor. “What happens when all of us leave because we can’t stand it here any more?” asks Sophie Jungnickel, a manager for a local business.

Plenty of the AfD’s foes seem inclined to quietly let Mr Siegmund mess up, hoping a mishap in a small state will remove the party’s sheen of invincibility. Residents of Saxony-Anhalt feel differently. “It’s easy to say these things in Berlin or southern Germany,” says Lennart Birth, a political activist who recently left the state. “But not for those who have to live with it every day.”■

ch089.xhtml

Cornered, Vladimir Putin plans to escalate his war

His refusal to back down is rooted in Russia’s gangster politics

Photograph: IMAGO

Aug 30th 2026

On August 29th at a university in Moscow, Vladimir Putin observed that orcas feed on polar bears. “That’s the food chain,” Russia’s president stated. “Children should be taught about this, so that they understand what kind of world we live in and why we are conducting the special military operation”—ie, Russia’s war on Ukraine. The claim was fanciful: there is no known case of an orca eating a polar bear. But it provided insight into the Russian president’s predatory worldview.

Recently NATO countries have become increasingly concerned about the direction Mr Putin’s thoughts are taking. Russia’s army has been largely stalemated for years. The war has become a contest of drone and missile attacks that, while devastating to Ukraine, are damaging Russia too. But Mr Putin is not backing off. The worry is that he may try to escape from a tupik (dead end) as he has in the past: by finding some new way to escalate the conflict.

On August 25th John Ratcliffe, the director of America’s CIA, unexpectedly flew to Moscow. Accounts differ on whether he went to warn Mr Putin not to attack NATO countries, or to deliver a bleak assessment of Russia’s prospects and persuade him to negotiate. Three days later Finland requested the help of Swedish aircraft and warships in monitoring its border with Russia for incursions. On September 1st Germany openly blamed Russia for a drone attack at Leipzig’s airport on August 4th, and imposed some trifling sanctions. In late August Romania destroyed a Russian naval drone that targeted a drilling platform in its waters, and Slovakia foiled an arson attack on a drone factory.

Meanwhile Mr Putin has ended his silence over Ukraine’s drone attacks on Russian refineries and e-retailer warehouses. In an address on August 22nd he accused Ukraine of opening a “Pandora’s box” and promised “blowback on the most sensitive sectors of your economy”. Earlier that day he dismissed any notion of ending the war: “The special military operation continues.”

Sources close to the Kremlin say that, after several weeks of consideration over the summer, he has chosen to escalate. Indeed, a longtime associate of Mr Putin says he is likely to have seen Mr Ratcliffe’s visit as a sign of American weakness. That, the source said, was the president’s interpretation of the visit to Moscow in November 2021 by Bill Burns, Joe Biden’s CIA director, to warn him against invading Ukraine.

No one can know what is inside Mr Putin’s head. But his reasons for escalating, like the decision to invade, are internal. His key objective has always been preserving his own power, and ultimately his life.

One concern for the Kremlin is the evaporation of the public’s silent assent for the war. Deep strikes by Ukrainian drones, along with internet restrictions, have shattered the veneer of normal life. The Levada Centre, an independent pollster, says 57% of Russians describe the situation as “tense”; respondents cite the war, drone attacks and the lack of a sense of the future. Mr Putin’s approval rating is sliding.

The mood among elites is even more sour. The biggest issue, one insider says, is the time the war has wasted: whatever the outcome, nearly five years is a failure. There is a growing feeling that the tsar is naked, says another. None of Russia’s problems is fatal to Mr Putin: he has the economic resources and the repressive apparatus to carry on. But perception matters.

Mr Putin may express his aims in the language of geopolitics, but in fact he is guided by criminal logic: never show weakness. Russia’s system has a high tolerance for violence, but is unforgiving of failure. He may believe that failure to achieve even his minimum goal, taking all of Ukraine’s eastern Donbas region, would put him physically at risk. (“They will hang me,” he told cronies at one point during the war, according to an insider.) Simply carrying on is no solution. “Inertia-driven scenarios are disadvantageous,” one member of the Russian elite says: they consume economic resources and political capital.

It is not clear how Russia would escalate. It is likely to intensify hybrid warfare against Ukraine’s allies and target their supply routes to Ukraine. Western intelligence agencies worry that Mr Putin might target factories in Europe that make arms for the war effort. But they think he does not want direct confrontation with NATO.

Within Ukraine, escalation largely means scaling up things Russia is already doing. On August 30th Russia’s Ministry of Defence said it was preparing a massive strike on Ukraine’s energy infrastructure. But Ukraine already expects such attacks. As for the ground war, Russian sources say Mr Putin is committed to throwing an additional 300,000-400,000 men into the meat grinder. Extra men will probably be recruited by regional governments to keep any political discontent localised. The authorities are already testing the new recruitment system in Penza, a region 750km south-east of Moscow.

Mr Putin is also intensifying repression of supposed “internal enemies”, not because they pose any threat but to demonstrate strength. On September 2nd police arrested Lidia Moniava, the founder of a hospice for terminally ill children. She faces a 10-year jail sentence for an antiwar social-media post in 2023. Two days earlier authorities dropped all charges against Ilya Traber, a businessman in St Petersburg, citing his ill health. Mr Traber, who has long ties to Mr Putin, had been charged with a contract killing. ■

ch090.xhtml

Haakon VIII, Norway’s new king, has a big crown to fill

His father was loved by royalists and republicans alike

Photograph: Reuters

Sep 3rd 2026

NORWay is not much for pomp. But when Haakon VIII swore the oath of allegiance before parliament on September 1st, some splendour was in order. The new king arrived in the Lincoln Continental that his father, Harald V, used for his wedding in 1968. Dressed in uniform next to Ingrid Alexandra, the crown princess, Haakon vowed to have the same good relationship with parliament that his father did. MPs sang the Kongesangen, the royal anthem, to the tune of “God Save the King”.

Despite their egalitarian mores, Norwegians broadly support the monarchy. Unusually, it has been endorsed in a referendum: in 1905, after Norway’s union with Sweden dissolved, 79% of voters plumped for a king over a republic. After Harald died on August 28th, aged 89, a poll by Norstat found 72% of Norwegians support having a king.

That is largely to Harald’s credit. Despite his Oxford degree and weakness for sailing yachts, the old king cultivated a regular-Jo air. His wife, Sonja, was the daughter of textile merchants. He praised Norway’s leftist partisans for their role resisting the Nazis, and spoke warmly of refugees. Klassekampen (Class Struggle), a socialist newspaper, put Harald’s obituary on its cover. The headline: “Also the republicans’ king”.

Haakon spent years as regent while his father was ill, but taking over will be tough. His family’s scandals make Britain’s royal peccadilloes seem less remarkable. The new queen, Mette-Marit, was friendly with Jeffrey Epstein long after his conviction for soliciting sex from a child (she has apologised). Her son from a previous relationship was convicted of rape in June. Other royals are merely barmy. Princess Martha Louise, Haakon’s sister, once ran a school that taught students how to talk to angels. Such affairs dragged support for the monarchy to just 60% in February.

Eventually Haakon may face a question his father avoided. As monarchs live longer, more are choosing to abdicate. Margarethe II of Denmark did so in 2024, and Emperor Akihito of Japan in 2019. The last three Dutch monarchs stepped down as they entered old age. But Haakon may balk, as his father did. In 1999 Harald described abdicating as an affront to tradition; he would only go if he went “completely bonkers”.

Parliament is considering letting other royals step in if both monarch and regent are unavailable. The idea is that this will help prevent cock-ups, as in April 2024 when a health minister’s dismissal was delayed because Harald was ill and Haakon was abroad. But the new king is only 53. MPs have plenty of time to deal with such problems. ■

ch091.xhtml

EU enlargement hits an Icelandic snag

A no vote on joining the EU was more about fish than geopolitics

Photograph: Getty Images

Aug 30th 2026 | Brussels

THE FIRST week of September is when European officialdom returns from a summer spent at the beach, rested and raring to take on the continent’s challenges. Alas, the mood of the rentrée has been marred by an unexpected snub from the mid-Atlantic. In a referendum on August 29th Icelanders voted narrowly not to reopen negotiations to join the European Union. Though the vote hinged largely on local concerns, it was an unwelcome back-to-work surprise for Brussels types. For a club accustomed to discussing how quickly others might join, being told “no thanks” is discomfiting.

Outsiders might have expected Iceland’s referendum to turn on the great geopolitical questions of the age. War is raging in Europe and Iceland’s neighbour Greenland is coveted by Donald Trump, the American president whose ambassador in Reykjavik once joked that the two islands could become America’s 51st and 52nd states. Most countries that can join alliances to protect themselves are doing so. Sweden and Finland became members of NATO soon after Russia invaded Ukraine. Ukraine has applied to join both NATO and the EU; and Canada—another country Mr Trump imagines as a future American state—is trying to get closer to the bloc. Iceland, which is already a NATO member, might seem an obvious candidate for the EU as well.

But all referendums are local. Iceland’s fishing industry, which accounts for almost 40% of the country’s goods exports, led the charge against EU membership. Fishing interests had derailed Iceland’s previous plan to join the union in 2013, after the country had suffered a crippling financial crisis. This time they made the same argument: the union’s rules would allow other countries’ vessels into Icelandic waters and impose caps on fishing that would hurt exports. The movement against joining was bolstered by a belief that Iceland can prosper on its own, and that existing arrangements with the EU—along with the country’s membership in NATO—are sufficient.

It wasn’t just the fish. Only inside Reykjavik, the capital, did most voters back restarting negotiations to join the EU; everywhere else, there was a clear majority against. Farmers fretted about competition and loss of government subsidies. Employers and some unions stood against it, too. That turned out to be more convincing than the campaign in favour, which claimed that a tiny island nation needs formal membership in a wider union to thrive.

The EU will no doubt play down the snub. Iceland has only 400,000 people—just 0.1% of the EU population, fewer than any current member. It is already in the European Economic Area, a sort of outer tier of the EU, and forms part of Schengen, the bloc’s passport-free travel scheme. The referendum was merely to reopen the negotiations that were suspended in 2013, after a change of government. The renewed talks might have failed anyway.

But this will be scant consolation for a union that has prided itself on its attractiveness to outsiders. Ten countries with a combined population of 150m have formally applied to join the union (though talks with Turkey, the largest, have been frozen for years). All of them are poor, most have dysfunctional politics and one (Ukraine) is at war. This weekend’s vote is a reminder that richer potential members of the EU, such as Switzerland and Norway, are in no rush to join.

It is a point of pride in Brussels that while Mr Trump threatens his neighbours with takeovers, the EU has such appeal that neighbouring countries are clamouring to join. Someone forgot to tell the Icelanders.■

ch092.xhtml

Tiny Montenegro still hopes it will be the EU’s next member

EU members, especially France, may be in no mood for enlargement

Photograph: Getty Images

Sep 3rd 2026

Western Balkan countries have a joke about their long efforts to join the European Union: “We pretend to reform and they pretend to want us.” Yet Russia’s invasion of Ukraine in 2022 injected new life into the process, and Montenegro’s government thinks it has a chance soon. With just 630,000 residents, Montenegro would cost the EU relatively little in aid. Its agricultural exports are too small to worry Europe’s farmers, and it poses no risk of significant labour migration.

At a summit in Montenegro in June, EU leaders fell over themselves to praise their host’s progress. The country’s government aims to finish aligning with the bloc’s standards by the end of this year, and join in 2028. Ursula von der Leyen, the president of the European Commission, called that schedule “within reach”. But Iceland’s vote on August 29th against restarting its accession negotiations has dampened the mood. Some hoped a bid by a rich northern country would boost momentum. Now the aspirants will simply have to persuade Europeans that bringing in poorer south-eastern countries is in their interests.

Map: The Economist

Of the western Balkan countries, Montenegro is at the head of the pack. Albania is next, while Kosovo and Bosnia & Herzegovina are making little progress. North Macedonia is blocked by its hostile neighbour Bulgaria. Serbia, the biggest of the six, got a red light from eight members in July, mainly over its autocratic governance. On August 27th Montenegro’s prime minister, Milojko Spajic, said that two neighbouring countries (clearly meaning Serbia and Albania) were trying to slow the country’s accession, presumably out of jealousy.

After Yugoslavia broke up, Montenegro was dominated for decades by one party, which led it to independence and NATO membership. In 2020 that was replaced by a coalition including pro-Serb groups. Europhiles feared Serbia would run the country’s foreign policy. That has not happened; relations between the countries are poor. The prospect of EU accession has led parties to co-operate on reforms.

Sceptics say the changes are cosmetic, and note that EU leaders have stopped talking about Montenegro’s powerful drug-smuggling clans. Expansion-minded Europeans worry that if a tiny country cannot get in, Ukraine and others have no hope. The biggest risk is France. President Emmanuel Macron supports accession. But French law requires a referendum or three-fifths of both houses of parliament to approve the accession of a new member. According to Eurobarometer, just 35% of the French support expanding the union. And Mr Macron will leave office next year. He may be succeeded by Marine Le Pen of the populist-right National Rally, who firmly opposes enlargement.

When Mr Macron visited Montenegro in June he signed a series of political and business agreements, including one aimed at building a favourable image of Montenegro in France. But Jovana Marovic, a former minister of European affairs in Montenegro, is not optimistic. A majority in favour of ratifying accession for any future members, she says, “simply does not exist” in France’s parliament.

“This is an extremely unpopular thing to say,” Ms Marovic continues. “The atmosphere in Montenegro is almost euphoric, as if EU accession were already a done deal.” She thinks there is time for her country to build support for accession among EU members. For the moment, many are as wary of letting in new countries as the Icelanders were about joining. ■

ch093.xhtml

America has a whopping debt. But Europe has the whopping debt problem

The French-size problem at the heart of the EU

Photograph: Peter Schrank

Sep 3rd 2026

To be European is to experience fine cuisines, walkable cities and constant nagging about the many ways in which your economy is falling behind America’s. In recent years weighty reports bemoaning Europe’s lagging productivity have become a literary genre in their own right. So news in recent weeks that America’s government debt had passed $40trn, raising questions about how long bond-market types will keep funding its persistent budget shortfalls, have been greeted in some European quarters with a degree of Schadenfreude (or whatever the Greek equivalent is). This is a continent that knows a thing or two about spiralling debt crises, after all. How refreshing it might be if the tables were turned, and Europeans could be the ones making recommendations about how politicians in Washington should be running their economic affairs.

Indeed, Europe looks at first glance to be in markedly better fiscal shape than its transatlantic rival—a rare case in which economic data seemingly favour the old continent. The total debt pile of governments in the European Union amounts to a mere €15.7trn ($18.2trn), in an economy not much smaller than America’s (Europeans are poorer but there are more of them). Two decades ago both places had debt-to-GDP ratios of around 65%. Now America’s has swollen to over 120%, and the IMF thinks it will be north of 140% by 2031. Europe’s is at “just” 83%, and its future is essentially flat—still in vaguely prudent territory. But Europeans tempted to subject those stateside to sermons about economic management (does America need a Draghi report of its own?) should hold their fire. For though America has the bigger debt, Europe has the bigger debt problem. America’s advantages include the world’s reserve currency and a growing economy. Europe’s disadvantage can be summed up in one word: France.

Comparing European and American debt is trickier than the headline numbers suggest. America owes too much, no doubt, but Europe owes it in all the wrong places. When it comes to handling interest payments on trillions in federal bonds, America is stronger than the sum of its parts: the entire union stands behind every dollar owed. In Europe, by contrast, most borrowing is contracted by 27 member states, each of which must repay its own debt lest the union as a whole totter. That makes the EU (or at least the euro zone at its heart) only as strong as its weakest link—as Greece once showed when it threatened default. Lots of countries in the EU have manageable debt piles, notably Germany and its northern neighbours. Others, like Greece and Spain, have high debt but enough growth to appease the bond vigilantes. France stands out for having neither fiscal space (its debts amount to 118% of GDP) nor growth (an underwhelming 0% in the most recent quarter).

As a result, interest rates on French ten-year bonds have now reached 4.2%, their highest in 18 years—higher even than Italy, the continent’s traditional problem child. In part this is a global dynamic: governments everywhere are paying more interest, as investors fret over inflation and geopolitics. But the spread between French and German bonds, the extra interest France has to offer lenders to entice them to hold its riskier debt, is now close to one percentage point. The fragmentation of rates among EU countries is still nowhere near the levels seen during the euro-zone crisis of the 2010s. That is largely thanks to the European Central Bank, which has since signalled it might step in if some euro-zone government’s borrowing costs were to rise too high. But the central bank’s largesse is meant to be conditional on that struggling government agreeing to make a serious effort at enacting reforms and remedying its fiscal mess.

Politics tends to interfere with such sound policymaking, and Europe has a surfeit of politics coming its way. In 2027 national elections are due to be held in France, Greece, Italy and Spain, four debt-laden countries. Fiscal discipline will no doubt slip as voters are showered with goodies. France again stands out. The front-runner in the polls is Marine Le Pen, a populist who once campaigned to leave the euro. More recently, the head of her party suggested the ECB could be cajoled into helping in some way. Her opponent on the populist left, Jean-Luc Mélenchon, has announced he wants to “set fire” to a portion of France’s government bonds. Even among level-headed centrists, nobody is suggesting a return to deficits of 3%, which EU treaties mandate, any time soon.

Après moi, le debt-luge

Any suggestion that France is the next Greece is hyperbolic nonsense—at least for now. But French profligacy is reshaping Europe nonetheless. America can run large deficits in part because its vast stock of treasury bills are denominated in the world’s reserve currency, and are thus a useful haven for nervy investors in choppy times. Europe would love to capture some of this “exorbitant privilege”. One way it tried to do so was by getting the EU to issue its own federal bonds, jointly backed by all 27 countries. The bloc started doing so on a large scale in 2021, to finance its pandemic recovery plan. Many hoped to issue more such debt in future to finance defence spending, in keeping with calls by Emmanuel Macron, France’s president, for more European “strategic autonomy”.

But continued French budgetary drift has probably put paid to further joint borrowing. The bigger the spread between French and German bond yields, the bigger the implicit subsidy flowing from Berlin to Paris (and more generally from the fiscally flush to the skint across the EU) whenever joint bonds are issued. Germany, never keen on extending the idea in the first place, now has a sizeable argument against it: who would volunteer to run a joint tab with a spendthrift whose creditworthiness is doubted by markets? French fiscal incontinence has long had a cost for the French. Now other Europeans are paying for it, too. ■

ch094.xhtml

The changing geography of violent crime in Britain

Violent crime is falling fastest in cities like London. It’s the counties you should worry about

Illustration: Gregori Saavedra

Sep 3rd 2026

On January 28th 2022, 18-year-old Joe Dix was found lying in the street. He had been stabbed seven times by members of a rival drugs gang—one of them wielding a 24-inch (61cm) sword. Dix died an hour later. “His insides were coming out,” his partner would later tell the court.

Where do you think this happened? London, notorious for knife crime, is an obvious choice. Birmingham, Manchester and Liverpool, each with its own history of gang violence, are others. In fact Dix was murdered in Norwich, a small, quaint city the Sunday Times recently crowned the best place to live in Britain.

More than 70% of Britons believe knife attacks are rising across the country. They are not. Hospitalisations for “assault by sharp object”, one of the most reliable indicators of violent crime, have fallen by a third in England since 2019. Last year homicides fell to a 50-year low.

Social media’s doomsterism goes some way to explaining this disconnect, as do spikes in shoplifting and phone snatching. A less appreciated reason is that the geography of violent crime is changing. The urban centres historically associated with violence have achieved impressive reductions over the past decade. Rates have fallen much more slowly in the counties. In some they have increased.

Consider Birmingham, London and Manchester. In the year to March 2018, the start of Britain’s knife-crime epidemic, these three cities accounted for 47% of the country’s knife-related homicides and attempted murders. In the year ending March 2026 this was down to 36%. In each city, homicides have fallen by more than 30% from their peak. Progress has been much slower in places like Suffolk (down by 18%), Hampshire (14%), Nottinghamshire (13%) and Cambridgeshire (up 23%).

Chart: The Economist

Hospitalisations for “assault by sharp object” show a similar pattern. Most police forces have data only up to the year ending March 2025. Look at a two-year rolling average and London’s is down by 28% from the peak years of 2018 and 2019. Much more modest falls have taken place in Staffordshire (down by 14%), Durham (7%), Devon and Cornwall (5%) and Hertfordshire (5%). In Sussex there has not been any drop at all (see chart).

Hospitalisation rates are growing fastest, albeit from a low base, in Cambridgeshire and Bedfordshire—leafy, suburban communities that last year suffered increases of 60% and 50%, respectively. Middlesbrough, a poor town but not one that has a long history of severe street violence, is recording more sharp-object hospitalisations than it did during Britain’s peak years. Following the deaths of two police officers and seven other people, and growing public disorder, it has become the site of an organised-crime investigation. Even within London, police-recorded violence is growing fastest in boroughs such as Hillingdon (up by 9% last year) and Kingston upon Thames (up by 13%), which are relatively affluent areas on the outskirts. Overall the upsurge of the 2010s is proving most resilient outside the urban centres usually associated with violent crime.

This is largely the result of changes in the structure of Britain’s drugs market. Violent crime in Britain is closely linked to drug trafficking. This is as true in the counties as it is in cities. A study published in 2021 looking at six years’ worth of air-ambulance call-outs in Kent, for example, found that instances of “penetrating trauma” tend to occur along drug-trafficking routes. These “county lines” emerged in the late 2010s as the advent of the dark web (used more for wholesale drug purchases than retail), together with cuts in police numbers, led to a proliferation of drug gangs in urban centres. Competition between them was the main cause of Britain’s knife-crime surge in the late 2010s.

It also brought more violence to the counties, as gangs looked for less crowded markets outside the cities. They built a business model using vulnerable children as mules—ferrying deliveries from urban gangs to rural addicts. Knives are a way to protect yourself, and your product, from rivals who would take it.

Thankfully most areas are recording less violence than they did during this first wave of expansion. But rural violence is proving sticky because such areas remain the most profitable route to growth for drug-dealing groups. In 2020 a Home Office review found that the levelling-off of urban drug use was driving gangs towards new markets around towns and universities. According to annual threat assessments, the number of county lines known to the authorities has increased from 720 in 2017 to more than 6,500 today. The dealers have also expanded their product lines. While the original county-lines gangs dealt mainly in crack and heroin, they now also traffic in cannabis, powdered cocaine, ketamine and prescription drugs.

Increased competition in rural areas has created a market based on customer service. Above all that means speedy deliveries. In many British towns it is now possible to have drugs delivered within an hour—think Amazon Prime for amphetamines. That is a difficult service for a business based in London to provide for customers in Margate. Supply lines have had to adapt. “The classic county-lines model of sending your boys down on the train has gone,” explains Simon Harding, a criminologist. “It’s much easier to set up a local hub and deal from there.”

The result is that drug gangs are increasingly indigenous to rural areas, dealing from local properties and recruiting local children. The number of “internal lines” (those operating within rather than between counties) known to the police trebled between 2022 and 2025. The share of lines originating in “task-force areas” (covering the traditional hotspots) fell from 70% to 50%.

It is a shift the government appears unprepared for. The Knife Crime Action Plan, released at the start of April, says a lot of the right things. But it also adopts an outdated definition of the county-lines model—promising to empower the British Transport Police to disrupt the train networks used to move drugs from cities to towns, and to boost funding to County Lines Taskforces. This targets a business model that is already changing.

Knives outdated

When so much violence is perpetrated by vulnerable young people recruited by predatory organised criminals, prevention is a public-health problem as much as a policing one. In the traditional hotbeds of knife violence, local experience with gangs has produced a certain amount of understanding of criminal practices, from clued-up family members to streetwise youth workers. Rural areas lack that kind of exposure. Many are used to seeing knife crime as an urban, even a racial, problem best solved by tougher policing. “That is a very difficult mindset to get out of,” admits Niven Rennie, a former police officer who served as director of the Scottish Violence Reduction Unit. “But once you understand it, you don’t go back.” ■

ch095.xhtml

London’s approach to knife crime is working

Sadiq Khan has mainstreamed a public-health approach that was previously fringe

Illustration: Gregori Saavedra

Sep 3rd 2026

London is known for many things. Among the least flattering is its reputation as a hotbed for knife crime. Last year Mexico, a country at war with drug cartels, updated its travel guidance to warn about knife assaults in Britain’s capital. According to Donald Trump, Londoners live in fear of being “stabbed in the ass”.

This was never really true. For six of the past eight years, New Yorkers have been more likely to be stabbed to death than Londoners (and they have been at least ten times as likely to be killed by other means). That doesn’t mean the reputation was a complete fabrication. Knife crime rose dramatically in Britain from 2017, especially among teenagers. In the year to March 2018 more than 280 people were stabbed to death in England and Wales, up from the previous five-year average of just over 200 a year. London, the core of the crisis, accounted for 108 of those homicides (compared with 62 deadly stabbings in New York that year).

Since then rates of violence have declined precipitously. Hospitalisations for “assault by sharp object” have fallen by a third in London since their peak. Knife-related homicides have shrunk by a half. Last year teenage homicides in the capital fell to their joint lowest level (similar to 2012) in almost 30 years. In 2025 you were twice as likely to be stabbed to death in New York as in London.

Thank Sadiq Khan. Right-wingers like to blame London’s Labour mayor for the capital’s imagined lawlessness. In fact Lord Khan introduced England to a knife-crime strategy that appears to be working. It has since spread beyond London.

As English politicians scrambled for answers during the 2017-19 “epidemic”, Lord Khan looked to Scotland. In the previous decade Glasgow, once labelled Europe’s murder capital, had achieved impressive reductions in violent crime. One reason was the roll-out of the Scottish Violence Reduction Unit (VRU), which aims to prevent violent crime by bringing the police, schools, hospitals and sports clubs together to identify at-risk children and direct them away from crime.

Between 2008 and 2018 Scotland’s VRUs were credited with bringing about a 38% fall in homicides and a 43% drop in attempted murders and serious assaults—many of them knife-related. Lord Khan announced England’s first VRU in London in 2018. Today there are 20 such units across England and Wales, covering areas that account for 80% of all knife crime.

A recent Home Office review found that England’s VRUs caused a 12% fall, since 2019, in hospital admissions for violent assault among the under-25s. That is modest compared with Glasgow, but the Home Office looked only at the national average. Violent crime has fallen most in cities like London—dense urban areas where the model is easiest to implement.

There are caveats. Alcohol use, a big factor in violent crime, has dropped across the country, too. Scotland’s success, which occurred around the same time as changes to licensing laws, cannot be put down just to its VRU. Nor can London’s (between mid-2020 and mid-2024 London lost 29% of its nightclubs). It also takes time to work out what produces the best results in a given area. Scotland did well by placing youth workers in accident-and-emergency departments to reach children with violence-related injuries. VRUs in England and Wales, still in an earlier stage, are spending a lot on interventions, such as school visits, that show little evidence of working.

The overarching logic, however, is sound. A 2023 study by the Youth Endowment Fund, a charity, found that having two or more adverse childhood experiences increases the risk of gang involvement by 61%. In Britain today violent crime is largely driven by organised-crime groups relying on vulnerable young people to traffic drugs. That is different from early-2000s Glasgow, where territorial street gangs were much more interested in recreational violence and turf wars than in drug trafficking. But in either case VRUs can help restrict the labour market available to exploitative gangs.

Scotland’s success, according to Iain Brennan, a criminologist, was not just to adopt the model but to change the hearts and minds of all involved. “It’s the philosophy of violence reduction, of changing the mindset that is the most important.” Lord Khan is often ridiculed for what can appear a fluffy approach, emphasising the importance of youth clubs and keeping troublesome children at school. But his soft approach is having hard effects. ■

ch096.xhtml

Has London passed peak phone snatching?

There are promising signs

Illustration: Gregori Saavedra

Sep 3rd 2026

London is far safer than it was. But one crime category has exploded in recent years: theft. Both shoplifting and “theft from the person” (mostly phone snatching) have roughly doubled since 2020. In the year to July 2026, rates of shoplifting hit a new high. But London’s rolling 12-month total for personal theft has fallen by 27% from its peak in January 2025 (see chart).

Chart: The Economist

The fall appears to be due to improved police tactics. Phone snatchers specialise in drive-by theft—using high-powered electric bikes to silently approach, and speed away from, pedestrians on their phone. For the police, chasing such thieves has been a losing battle. Officers aim to be at emergency calls within 15 minutes. That is plenty of time for someone on a 50mph e-bike to make their escape, zipping down alleyways and through heavy traffic to outmanoeuvre chunky police vehicles.

Tactics have had to adapt. The Metropolitan Police has been building up its own e-bike stockpile to chase down suspects. It has also stepped up efforts to confiscate, en masse, those used by criminals. And since October it has deployed drones to follow thieves from a distance and pick them up once they let down their guard. Combined with targeted policing in the tourist-heavy areas where most phone snatching occurs (such as Westminster), these new strategies appear to be working.

Police officers have also got wise to the business model of phone snatching. What looks like petty theft in fact depends on an organised network of handlers exporting stolen phones abroad. So a few arrests higher up the chain can have an outsize effect. In October the Met arrested 18 people suspected of sending up to 40,000 phones to China (50% of London’s total).

Almost everyone in London knows somebody who has had their phone stolen. That goes a long way to explaining rising feelings of insecurity in the capital and beyond. Shoplifting remains a blight. But when it comes to phone snatching, at least, London appears to be on the right track. ■

ch097.xhtml

Can Andy Burnham extend his winning streak in London?

A by-election will test whether the prime minister is more than just a northern phenomenon

Photograph: AP

Sep 3rd 2026

It is the year of the by-election in British politics. In February disgruntled voters in Manchester hammered a nail in the coffin of Sir Keir Starmer’s premiership by electing a Green MP. In June the northern constituency of Makerfield elected Andy Burnham in a landslide, with a mandate to replace Sir Keir as prime minister. And in July, with Mr Burnham now at the helm, Greater Manchester elected a Labour mayor to replace him, confirming the prime minister’s electoral prowess.

Now Londoners have their turn. On September 1st Sir Keir announced he was resigning as an MP. This triggered a by-election in his London constituency of Holborn and St Pancras, which stretches from touristy Covent Garden to lively Camden Town and leafy Primrose Hill. The contest is an opportunity for Mr Burnham to show that he can extend his winning streak beyond his native north-west England and stop the advance of the Greens in Labour’s urban heartlands.

In normal times Holborn and St Pancras would be a safe Labour seat. Young, diverse and highly educated, the area has been represented by Labour since 1964. Between 2017 and 2024, however, Labour’s vote share fell from 70% to 49%. Andrew Feinstein, a left-wing pro-Palestine activist, won 19% of the vote as an independent candidate, drawing on the area’s large student and Muslim populations. The Green Party won 10%. Zack Polanski, the Greens’ outspoken leader, has announced that he will stand in the by-election.

A consolidated left-wing challenge could be a headache for Labour. In local elections in May the Greens won around 29% of the vote, while Mr Feinstein’s slate won a further 6%. Even so, Labour managed to win 38%, despite the elections coming at the nadir of Sir Keir’s premiership. Since then, polls suggest Mr Burnham has won back a substantial chunk of left-leaning defectors.

By-elections are unpredictable, especially with the excitement that Mr Polanski’s candidacy will generate. But despite the risks, the contest in Holborn and St Pancras is above all an opportunity for Mr Burnham. A solid Labour victory would wallop the already floundering Greens, quash the idea that the new PM’s appeal is limited to the north and prolong the glow of his first weeks in power. The honeymoon is not over yet.■

ch098.xhtml

Why the world’s best female footballer joined a tiny London team

At stake is whether a women’s team can flourish independently of a men’s franchise

Photograph: Alamy

Sep 3rd 2026

Alexia Putellas is known as la reina (the queen). In 14 years at Barcelona, the Spanish footballer won the Women’s Champions League four times and ten domestic league titles. She lifted the World Cup with Spain in 2023 and has twice received the Ballon d’Or Féminin for the world’s best player. This summer she swapped Barcelona for the Bromley-based London City Lionesses (LCL), a seven-year-old team with a small stadium. Why?

LCL was originally the women’s offshoot of Millwall, a club best known for its fans’ predilection for fighting. In 2019 the Millwall Lionesses opted for independence and adopted a new name. Four years later, the club was bought by Michele Kang, an American entrepreneur who also owns women’s teams in America and France. Since winning promotion last year to the Women’s Super League (WSL), the top tier of English women’s football, LCL has gone on a remarkable spending spree. As well as Ms Putellas, the club has signed a former England goalkeeper, Mary Earps, and other stars from Australia and France.

A rich overseas owner splashing their cash is a familiar story in English men’s football. But for every Manchester City or Chelsea, for which investment brought success, there is a Leicester City or Portsmouth, where overspending led to the brink of ruin. The WSL is keen to avoid that. For the 2026-27 season, which kicks off on September 4th, clubs must operate within new spending limits. Outlays on players and wages must not exceed 80% of revenue, plus an owner’s own-money contribution equivalent to 25% of revenue or up to £4m ($5.5m), whichever is higher. Clubs that fail to comply can be punished with fines and points deductions.

The new rules are designed to encourage financial sustainability, but they also risk entrenching the dominance of the existing giants. Arsenal and Chelsea, for example, who between them have won the WSL for eight of the past nine seasons, each recorded revenue of €25m (£21m, or $29m) in their most recent accounts. They will have much larger budgets for transfers and wages than their rivals.

Ms Kang is betting on the ability of her club’s commercial team to increase its revenue dramatically and offset the spending on new players. In 2024-25 its total revenue was just over £900,000. Ms Putellas’s annual salary alone has been reported at close to that figure. Although revenue will have risen sharply since that time—a bigger WSL broadcasting deal with Sky and the BBC began in 2025-26—so too will have LCL’s costs. The club posted a loss of £10.6m in 2024-25. Revenue may need to rise as much as 15-fold to remain compliant with the WSL’s rules.

LCL is the only club in the 14-member league that is not linked to a larger men’s team. Christina Philippou, an expert in football finance at the University of Portsmouth, believes this is an advantage: “London City have a better ability to leverage the female-focused brands in a way that some of the affiliated teams cannot, because of the association with their men’s sides.” She also believes that Ms Kang’s ownership of Washington Spirit in America and OL Lyonnes in France could boost sponsorship, thanks to an ability to offer deals that reach three different markets.

But LCL also lacks the infrastructure that can give other women’s teams a leg up. Arsenal’s women now play their home matches at the Emirates Stadium and achieved average attendances of over 33,000 in 2025-26. For now, LCL borrows the stadium owned by a third-tier men’s side, Bromley. It attracted average gates of 3,200 last season.

Alex Stewart, CEO of Analytics FC, a football consultancy, argues, “The only way to disrupt the established wsl hegemony is to buy your way in.” Ms Kang has certainly done that. It is possible that Ms Putellas will prove the catalyst that establishes LCL as a permanent contender in women’s football. On the field, she could push the team into Champions League qualification, unlocking a new source of prize money. Off it, her profile could enable sponsorship deals of the magnitude required to make the club financially compliant. But it is a bold gamble.■

ch099.xhtml

The British government turns to influencers to get its message out

But that is no substitute for clear policy

Photograph: Getty Images

Sep 3rd 2026

AnDY BURNHAM grew up drenched in newsprint: as a boy he worked part-time delivering newspapers, and his first job after leaving university was as a reporter (unpaid) on the Middleton Guardian. Since becoming prime minister, he has shown no loyalty to the dead-tree press. He has done more interviews with podcasts than with newspapers, and more TikTok stunts than press conferences.

Every effective leader uses the media of their day. Lord Palmerston, prime minister in the mid-19th century, cultivated so close a relationship with the Morning Post that he would dictate its editorials almost word for word. Harold Wilson in the 1960s had tickers showing international news updates installed in Downing Street as he brought government communications into the telecoms era. Margaret Thatcher did interviews with publications like Smash Hits, an irreverent pop magazine. Sir Tony Blair brought “spin doctors” into the heart of government to deal with the barrage of coverage demanded by 24-hour news.

No wonder, then, that Mr Burnham has put such effort into a slick digital offering. He has posted more than one TikTok video a day on average since becoming prime minister. His online clips include ranking classic pub snacks and “ask me anything” Q&A sessions. Bad news, too, is delivered on social media. Mr Burnham used a Facebook post to admit problems with a scheme to increase prison capacity by releasing some criminals early. Many commenters praised the prime minister’s willingness to be open with the public.

Then there are the influencers. Mr Burnham has talked vintage football shirts with Ellis Platten, a YouTuber. He has posed with Bemi Orojuogun (pictured), who goes by “Bus Aunty” on TikTok, where her clips of buses attract millions of views. And he has built an electrical circuit with Big Manny, a science graduate who posts informative videos on Instagram.

Rishi Sunak was the first prime minister to invite influencers to Number 10 as a way of reaching younger voters who have increasingly switched off from old media (under-35s watch less than an hour a day of broadcast television, according to Ofcom, a regulator, compared with five hours for those aged 65 or older). Sir Keir Starmer went further, holding influencer-only press conferences and taking “content creators” with him on official trips abroad.

The practice of treating influencers on a par with old-school media has been formalised within Whitehall. An internal government document seen by The Economist warns civil servants that “people often resist being told what to do” by the government, but adds: “If the advice or suggestion comes from a relatable influencer, it can feel less like a directive and more like a shared discovery or recommendation.” Influencers can be paid “several thousand pounds” per post for taking part in government ad campaigns, the document also reveals. “The government has a duty to engage with the public where they are,” a spokesman says.

Politics and social-media culture can be a good fit, according to Georgia Harrison, a former contestant on “Love Island”, a reality-TV show, who now campaigns for online safety and has been to Downing Street several times. “It’s about causing the drama,” she says. Abigail Foster, a personal-finance influencer, argues that online personalities can reach areas that newspapers cannot: “We are where the next generation are, we look and sound like them.”

But jumping on social-media trends is no guarantee of popularity. Mr Sunak was unceremoniously dumped by the electorate in 2024, and Sir Keir’s whizzy videos did not shift his stiff image. “Keir is probably more Substack than TikTok,” admits James Lyons, who was his director of communications in Number 10 and is now senior partner at Penta Group, a consultancy. Ms Foster says she was asked to collaborate with Whitehall departments so often she started turning them down—and warns that the novelty of influencers filming inside Downing Street has worn off for their audiences.

Britain’s newspapers remain influential, often determining what is covered on the mighty state broadcaster, the BBC, and discussed online. Still, a political strategy which does not take social media seriously is doomed to fail. The internet allows “hyper-personalisation”, says Alexander Dragonetti, a former diplomat who appeared on “The Traitors”, a BBC game show, and now makes videos explaining foreign policy; politicians must explore a range of niche audiences in order to ensure they are reaching most voters.

The medium and the message

British politicians are still lagging behind their counterparts in America, where it has become normal for candidates to sit down for multi-hour podcast interviews. But so far, Mr Burnham is getting plaudits that his predecessor never did for his relaxed, mildly blokey online persona—even though most of his communications team also worked for Sir Keir.

Ultimately, the message and the messenger matter more than the medium. It is not coincidental that the two prime ministers in living memory who have best mastered the press, Thatcher and Sir Tony, were the two who had the best idea of what they wanted to achieve and how to go about it. Leaders who obsess over the minutiae of media coverage are too easily distracted. Influencers are quick to stress they cannot offer a magic bullet to struggling governments. As Ms Harrison says, “Policy has to come first and then communication, not the other way around.” ■

ch100.xhtml

France is love-bombing Britain

And Britain seems keen to return the compliment, for a change

Photograph: Getty Images

Sep 1st 2026 | LONDON AND PARIS

There are few more disconcerting experiences for the British than being love-bombed by the French. Britain’s neighbours across the English Channel are usually regarded as pesky rivals, exasperating partners and irritatingly well-tailored to boot. Yet this week France went out of its way to charm Britain.

On September 2nd President Emmanuel Macron and King Charles met at a high-profile event at the British Museum to inaugurate the Bayeux tapestry exhibition, set to be this season’s London blockbuster. It is the first time the 11th-century work depicting the Norman conquest of England in 1066 has left France. The loan “took probably more years” to negotiate than Brexit, Mr Macron joked last year; he has made it a personal centrepiece of bilateral diplomacy. “When we come together around one table,” he said of Anglo-French ties after announcing the exhibition in 2025, “then everything is possible.”

Why is France being so nice to Britain? Historical rivals, the two countries have sparred in recent years: over Brexit, submarine sales, small boats and more. When asked whether Mr Macron was a friend or foe, Liz Truss, fleetingly Britain’s prime minister, once said, “The jury’s out.” Now Britain seems keen to return the compliment. Under Sir Keir Starmer, Mr Macron got a state visit, a banquet at Windsor Castle and a partner with whom to build a “coalition of the willing” for Ukraine. Mr Macron was due on September 3rd to be the first foreign leader to visit Andy Burnham, Sir Keir’s successor, in Downing Street. The pair were expected to discuss Ukraine, security and Britain’s ties to the EU.

One clue to the new entente may in fact lie not in the British Museum but in a superb new French two-part biopic of Charles de Gaulle, “La Bataille de Gaulle”. Directed by Antonin Baudry, an ex-diplomat, it is—unusually for a French production—based on a British biography, “A Certain Idea of France: The Life of Charles de Gaulle”. Written by Julian Jackson, a historian at Queen Mary University of London, the 928-pager is the product of years of painstaking archival research.

A certain idea of les anglais © Getty Images

The big-screen diptych, starring Simon Abkarian in the title role and Sir Simon Russell Beale as a wonderfully crabby Winston Churchill, is naturellement a stirring tale of courage and moral clarity. This is the myth that to this day surrounds the résistant and founder of modern France. But, perhaps thanks to the foreign biographer’s eye, it is also an even-handed examination of both the visionary greatness and human fragility of le général.

Of particular interest for British viewers, the first part, “L’âge de fer”, dwells on De Gaulle’s years in exile in Hampstead, in north London, after the fall of France to Nazi Germany in 1940. It does not gloss over the high-minded general’s stormy relationship with Churchill. “Betting on De Gaulle seems something of a leap of faith, don’t you think?” Anthony Eden, the secretary of state for war, asks the prime minister. “This is not my first war,” Churchill retorts, adding clinically: “If he fails, we’ll get rid of him.” The general and the prime minister trade plenty of insults. But they also share moments of mutual understanding, admiration—and a flicker of affection.

London, after all, provided De Gaulle with refuge, offices (4 Carlton Gardens), a microphone (the BBC), backing (most of the time) and legitimacy (which the Americans resisted). The French have not forgotten. In 2020 Mr Macron awarded London, “totally exceptionally”, the Croix de la Légion d’honneur, France’s highest honour, as a mark of “eternal gratitude”. London, he said, was “the last bastion of hope when all seemed lost”.

What Mr Macron seems to be saying is that despite Brexit, despite Donald Trump, or perhaps because of both, France and Britain need to stick together now, as they did then. The De Gaulle biopic, which has drawn over 5m cinema-goers in France, not only evokes nostalgia for an age of unimaginable bravery. It also speaks to modern worries about national sovereignty, dependence and democratic values.

Stubbornly independent-minded, De Gaulle was infuriated by Roosevelt’s post-war plans for American military government of France, kept secret from him. The film includes the moment Churchill told the general gruffly that Britain would always choose America over Europe. Yet the doctrine of autonomy embodied by De Gaulle finds new resonance now that Mr Trump’s America is threatening to leave Europe to fend for itself. Despite the populist-right Reform UK party breathing down his neck, Mr Burnham has said that Britain needs to “be bolder” in getting closer to the EU. Mr Trump has already helped push Britain and France closer together. Last year the two countries reinforced co-operation over nuclear deterrence. The pair co-lead the “coalition of the willing”.

The film carries another message too: that the dark forces of nationalism can be beaten. The populist right is on the rise in both countries. Marine Le Pen is the current favourite to win next year’s French presidential election. On September 4th Jordan Bardella, her lieutenant, is due to be a guest of Nigel Farage, Reform’s leader, at the party’s conference in Birmingham. Mr Macron likes to urge the French not to cede to the “spirit of defeat”, a nod to a book about the fall of France in 1940 by Marc Bloch, a résistant gunned down by the Gestapo. The historical parallel may be imperfect. But this battle is yet another that Britain and France are in together. ■

ch101.xhtml

How work makes families more equal

The love of money is the root of quite a lot of good

Photograph: Reuters

Sep 3rd 2026 | Lucknow, Riyadh and Stockholm

Adecade ago Saudi women were barred from all but a few jobs. Then Saudi Arabia’s de facto ruler, Crown Prince Muhammad bin Salman, ended most restrictions—and even allowed Saudi women to drive. His aim was for the kingdom to rely less on oil and more on toil. Letting a previously housebound half of the population work was an obvious place to start.

This had dramatic consequences. Hamda, a young woman in Riyadh, heard the news with joy and told her husband she wanted to get a job. He forbade it. If she worked outside the home, she might talk to a man and bring shame on the family.

She argued with him for a year. Eventually, she won, for two reasons. First, extra money comes in handy. “I told him, we need to pay our bills,” she recalls. Second, her husband’s sisters all got jobs. Suddenly it became harder for him to claim, at the family dinner table, that only immoral women worked. So Hamda found a job in a mall, where she sells pyjamas. Her life is much better, she says. One day, she wants to buy a car and start her own business.

In much of the world, the big battle for women is not the one for equal pay, but the one to be allowed to work outside the home at all. In swathes of the Middle East, north Africa and South Asia, “male honour depends to a large degree on female seclusion,” notes Alice Evans of Stanford University. Fathers, brothers and husbands insist that for a woman to venture outside the home upends the natural order and puts her chastity at risk.

Such norms keep whole regions needlessly poor. In South Asia only a third of women are in the labour force; in the Middle East and north Africa, it is only a fifth. (In North America, by contrast, seven out of ten women work.) Removing the barriers to women working would raise per-person income by a fifth in many countries, estimates the World Bank—a bigger economic benefit than avoiding a civil war.

And there is more. As Hamda’s story illustrates, when women who were previously forced to stay at home start earning, the balance of power in the household shifts. Work makes families more equal.

The Economist crunched some numbers to explore the relationship between sexist norms and female labour. We used a measure called the “patrilineal/fraternal syndrome”, developed by Valerie Hudson of Texas A&M and Donna Lee Bowen and Perpetua Lynne Nielsen of Brigham Young University. It captures extreme chauvinism: things like the unequal treatment of women in family law and property rights, early marriage for girls, polygamy, son preference and tolerance of violence against women. We combined this measure with data from Demographic and Health Surveys (DHS), which poll women in low- and middle-income countries.

Chart: The Economist

We found a strong relationship (see chart). In the 21 most sexist countries, only 30% of women work, and the rate stays low regardless of how poor the household or the country is. As sexism recedes, more women work, especially in poor places, where the income is most needed. In the poorest households in poor countries with a middling score for sexism, the share rises to 55%, and to 80% in the least sexist ones.

Next, we looked at whether earning a wage affected women’s status at home. We used the Survey-based Women’s emPowERment index (SWPER), devised by Fernanda Ewerling and Aluisio Barros of the Federal University of Pelotas in Brazil, and others. This uses DHS data to assess how much decision-making power women have. Can they decide how the family’s budget is spent? Can they go out to visit friends or relatives without asking permission? In short, do their voices count?

These results were striking, too (see chart). Even after controlling for how sexist a country is and how rich a household is, women with jobs consistently reported having more say at home.

There are various possible reasons for this. In societies where women are expected to be submissive, going out to work may teach them how to be more assertive—as Hamda’s growing confidence illustrates. Also, when women earn a big chunk of the household income, it is easier for them to have a say in how it is spent. Finally, when women can be financially independent, they can leave an unsatisfactory relationship, or credibly threaten to do so. So husbands must treat them better or risk ending up single. “Money is power,” says Sakine Madon, a Kurdish-Swedish writer. “As soon as women get more independent, they can divorce from bad marriages.”

Defying convention

In some places the obstacles to women working are formal and legal. According to the World Bank’s index on “Women, Business and the Law”, some 540m women live in countries where legal equality remains “distant”. Penny Goldberg of Yale University finds that a country’s score on this index is, unsurprisingly, “highly correlated” with how well women do in the workplace.

Thus, changing laws can help, as Saudi Arabia shows. But it is only a start. Social norms are often a deeper challenge, notes Kathleen Beegle of the World Bank. Things “like which household member is responsible for pounding the foo-foo in the morning and making sure that dinner is on the table at 5.30pm; it’s not something that is really in the traditional policy sphere for governments.”

In a conservative neighbourhood or village, the first women to reject old norms may face intense pressure to shut up and conform. Consider Sandhya, a community worker in Uttar Pradesh in northern India. She grew up in a rural Hindu village where “men never go in the kitchen.” She was married at 15, to a man who forbade her to go anywhere without a male relative. He beat her when she talked back.

After six years, she left him and went back to live with her parents. Wanting to support herself and her son, she started working, initially as a nurse. Poisonous gossip spread. Why was she coming home late, wearing nice clothes? Was she a prostitute? The slander did not kill her ambition. But it will make other local women think twice before seeking independence.

Chart: The Economist

In rural India, as in many places, gossip helps preserve the status quo. Men routinely find higher wages by migrating to a city, but any woman who tries to do the same may be tarred as a sex worker, making it hard for her to marry. So nearly all rural-urban migrant workers in India are male. That makes the places where migrants lodge scary for women, and gives men cause to object to a daughter leaving the village—barring her way for her protection, they say. If this means she is around to cook and clean, they see that as a bonus.

In countries with deep-seated sexist norms, the technocratic toolkit for getting women into jobs is less effective. Subsidised child care may do wonders for Nordic women. But in Egypt, when an experiment offered mothers vouchers covering some or all of the cost of a creche for their children, only 11% took up the offer. Chauvinist hubbies outweigh cheap nurseries.

Yet norms change. Money, technology, social contagion and moving all can help, finds a World Bank study by Daniel Halim, Michael O’Sullivan and Abhilasha Sahay.

First, money. When a woman goes out to work, male relatives may feel that their honour has been sullied. However, even reactionary men like extra moolah. Dr Evans calls this the “honour/income trade-off”.

Different societies have dealt with it differently. China has eagerly chosen income. (The communist revolution set out to smash patriarchal traditions by pushing women into the workplace; then capitalist reforms raised the rewards for families whose daughters went to work in factories.) This is perhaps why, although China’s working-age female population is only marginally larger than India’s, its female labour force is nearly three times larger.

Cash persuades. When a woman earns a decent wage, her menfolk may have to reconsider their belief that women are not capable of such things. “My own brother didn’t believe in me in the past,” says Wafa al-Otiby, a beauty-salon owner in Riyadh. She asked him for a loan to start a home makeup studio. He lent her some cash, but laughed at her, telling her she did not know what she was doing and adding that in any case it was shameful. But when she paid him back within six months, he asked if he could be a partner in her thriving business. “Men’s attitudes toward women have changed,” says Ms Otiby. “They’re surprised now at what women can do.”

Wages motivate women, too, especially if they keep control of them. A study by Erica Field of Duke University and others found that Indian women on a public-works scheme worked more if their pay was deposited into accounts of their own, instead of those of their husbands’, and if they were trained in how to use them.

Dough changes attitudes among the better-off, too. Parnita Agarwal, the daughter of a small-factory owner in Uttar Pradesh, recalls growing up aware of families where husbands were bullies and wives were expected to be home by 7 or 8pm. She vowed to earn her own money, so that “if I’m ever stuck in an abusive marriage, [I can] afford my own housing and not think twice before leaving.”

When she won a place at Manchester University in Britain, some of her relatives told her parents not to let her go. The money would be better spent on a dowry, they said, since a girl’s aim should be to get married. Besides, studying abroad would make her impossible to control, they warned.

She went anyway. And when she came back to India, brimming with ideas about how to boost the family business and enrich her parents, her previously antagonistic relatives realised that they had been wrong. Now they are sending their daughters off to study, too, says Ms Agarwal.

Photograph: Getty Images

Connectivity accelerates change. In a very sexist society, digital technology enables women “to envision a different life”, writes Roya Mahboob, an Afghan tech guru. Often the most persuasive examples come not from the most liberal countries, which may be far off and alien, but from slightly less conservative neighbours. Nawal al-Onazi, an HR manager in Saudi Arabia, says she was inspired by satellite television from other Arab countries when she was young. She saw interviews with women with important jobs, and Egyptian comedies that showed less rigid forms of family life. She ended up divorcing a husband who wouldn’t let her work.

Technology can sway families. A study in Uttar Pradesh by Madeline McKelway of Dartmouth College finds that showing husbands and in-laws a six-minute video of women talking about their experience on a job-training programme made it 35-57% more likely that their wives would get jobs. The video stressed that the workplace (a weaving factory) was safe and all-female and the women working there still found time for household chores.

Technology can also help women pursue specific job opportunities. When vacancies are listed online, women can browse and apply without leaving home. Some can work remotely, too. A study by Jing Lu of Zhejiang University found that by making job-searching easier and creating new female-friendly occupations such as online customer service, the spread of digital technology increased the likelihood of women in China having jobs.

Then there is contagion. People are social creatures. Rather than resolving moral questions from first principles, most take cues from their peers. A study in 2020 by Leonardo Bursztyn of the University of Chicago found that most young married Saudi men thought women should be allowed to work outside the home. However, three-quarters underestimated how many other Saudi men agreed with them. When told the true number, they were more likely to help their wives find jobs.

Ms Otiby, the beauty-salon owner in Riyadh, says that even traditional Saudi men have now “started seeing women among their colleagues…and it has changed people’s thinking about marriage, too. It’s amazing! People now are getting married after meeting one another through work, instead of the traditional way.”

The kind of work that women do matters, suggests Johanna Mollerstrom of George Mason University. If most of the jobs available to married women are unsafe and unclean, a working wife is a signal to others (say, in the neighbourhood or church) that her able-bodied husband is lazy. As a country becomes richer and those jobs become nicer, the stigma fades.

Dr Evans puts it differently. The “honour/income trade-off” can change if the opportunity cost of women not working rises (because job opportunities improve) or the perceived cost (in “dishonour”) of women working falls. Saudi Arabia’s leader has been pulling both these levers, with quotas for firms hiring Saudi nationals and loud signals that “it’s respectable for women to be in the public sphere”, such as royal women doing fashion shows.

Slipping surly bonds

Perhaps the most dramatic way to escape from patriarchal norms is to move somewhere where they no longer apply. When Sara Mohammad was 17, her brother held a Kalashnikov to her head, demanding that she agree to an “exchange marriage”. He wanted her to wed a man twice her age whom she had never met, and in exchange her brother would be allowed to marry the man’s sister. Ms Mohammad, who was then living in Iraq, said no, twice. Her brother threatened that if she said “no” a third time, he would kill her, “as easily as drinking a glass of water”. So she said yes.

Then, as the wedding was being prepared, she slipped away under a borrowed burqa, took refuge with a sympathetic cousin and eventually found asylum in Sweden. Now she runs GAPF, a charity that campaigns against honour-based violence—and has a partner whose permission she does not need to ask to do things.

The process by which migrants abandon old customs is not straightforward, however. Women rarely migrate on their own. Usually they arrive with male relatives, who may be keen to isolate them from the host society. And the arrival of people from honour cultures in liberal countries is often unpopular. Many Swedes, for example, think the surge of asylum-seekers from Muslim countries a decade ago will undermine local feminist norms. This is a potent rallying cry for the populist right in Europe and America.

Yet such fears are overblown. Natives seldom adopt the attitudes of recent arrivals, whereas immigrants tend to assimilate. In America immigrants from countries where few women work have nearly closed the employment gap with native women by the second generation. A study of immigrants to 18 European countries from 46 other places found that female second-generation migrants hold more egalitarian views than male ones, and that the more unequal their country of ancestry, the larger the relative shift among women.

Liberal cultures are attractive, especially for women. Kinem (not her real name), who moved from Turkey to Sweden as a girl, decided that she did not want the “copy and paste” lifestyle of her female cousins, who “didn’t study”, married “quite young” and “all had the same apartments with the same furniture”.

Her rebellion came in stages. Her parents were not the strictest; they let her go to university. But they wanted her to come home every weekend. She stopped visiting so often, and started going to parties, though she did not drink alcohol. With a degree in chemical engineering, she found a series of good jobs in medical technology and bought her own apartment in Stockholm. That bothered her parents, but she held firm. Finally, she married a Swede. That was too much for her parents, who cut off ties with her for five years.

Eventually they were reconciled. They now accept that she will make her own choices. Today, she says, roughly half of her female cousins have jobs. “They have realised that they can also bring in money to the family. And people like money.” ■

ch102.xhtml

The Ukraine war is intensifying, expanding and stuck

The bleak view from Sweden, a Ukrainian ally braced for 50 years of confrontation with Russia

Photograph: Chloe Cushman

Sep 1st 2026

AWEARY world may not want to hear it, but the war in Ukraine is intensifying, not ending. As it deepens, the conflict is also expanding. Increasingly, serious-minded governments across northern Europe are planning to defend themselves against Russia for decades to come.

Prosperous, prudent Sweden is a case in point. It is only a couple of years since the Nordic country joined NATO, driven by Mr Putin’s aggression to abandon a two-centuries-old policy of neutrality, bordering at times on wide-eyed pacifism. As Swedes head to the polls on September 13th, support for Ukraine is such a consensus that the war has barely featured in debates. All major parties, including those on the populist right and left that once flirted with sympathy for Russia, endorse financial and military help for Ukraine, deep distrust of President Vladimir Putin and the need for strong national defences.

In Stockholm it is striking how often senior figures use the word “naive” to describe what was their country’s approach to Russia as recently as the late 2000s. Bigwigs on the right and left are at pains to show how coldly realistic they now are.

Under its current, conservative coalition government, Sweden has been outspoken in calling Russia weaker than it pretends to be, accusing Mr Putin and his aides of using dodgy statistics to hide the parlous state of its finances while using repression to silence public dissent. Yet the prime minister, Ulf Kristersson, does not claim that Mr Putin’s war-fever is about to break. Instead, in an interview on August 27th for “Inside Geopolitics”, a video show produced by The Economist, Mr Kristersson says he sees “no signs at all that Russia has changed its basic logic behind this war. They want to conquer Ukraine basically, and we don’t know what kind of aims or goals they have after that.”

Mr Kristersson concedes that Ukraine is feeling increasing pain, too, as Russian missiles and jet-powered drones bring terror to cities left vulnerable by a growing shortage of interceptor missiles—a problem that he notes has been made worse by war in the Middle East, which has burned through stocks. But in the next breath, he warns Russia not to expect Ukrainian resistance to crack, calling the country extremely resilient.

Seen from Stockholm, the collision of two unstoppable forces—Russian neo-imperialist ambition and Ukrainian patriotism—has generated a stalemate. The costs of conflict are rising, but both sides consider defeat still more agonising. The prime minister admits that he has no idea when the war might end. “But how it ends will define the security situation in our part of the world for at least a generation to come,” he says.

To prepare, his government has sought to bind Sweden to Ukraine as a long-term defence and industrial partner, donating and selling Gripen fighter-jets to Ukraine’s air force. It is expanding military conscription and deepening defence and security ties with its Nordic and Baltic neighbours, as well as with Britain, France, Germany, the Netherlands and Poland.

The line on Russia is every bit as tough on the centre left, which polls suggest is likely to win this month’s elections. Leading figures in the Social Democratic Party, which spent the cold war loudly opposing NATO membership, now describe a consensus across the political spectrum that Russia poses a “long-term existential threat” for the foreseeable future, requiring Sweden to make a 50-year commitment to strengthening its national defences, both military and social.

In Stockholm an end to Ukraine’s war is both longed for and feared. Senior figures worry that European unity, which has hitherto held up impressively, will be challenged as countries farther away from Russia demand warmer, more normal ties with Mr Putin. The foreign-affairs spokesman for the Social Democrats and potential future foreign minister, Morgan Johansson, worries that a ceasefire will lead only to a frozen conflict in Ukraine, not a durable peace. Asked whether Sweden will ever be able to trust Russia, he imagines a far-off future in which a change of regime brings to power a Gorbachev-like reformer. Even then, says Mr Johansson, Swedes “would have to keep our guard up” and avoid mistakes made 25 years ago, when political leaders stopped fearing Russian aggression and shrank the country’s defences.

Sweden is one of the most vocal advocates for Ukraine to be given a credible path to EU membership, as soon as Ukraine can show progress on corruption and governance, to anchor it to the West and shore up its democracy. Other big states, such as France, are backers of Ukraine but wary of EU enlargement. Some EU members with lots of farmers worry about competition from Ukraine’s arable industry.

No optimism, no alternatives

These differences make Sweden an unusually good place to ponder the fragility of Western unity around the war, especially if Donald Trump tries to force Ukraine to settle for an unfair, insecure peace. Sweden has sought to maintain good relations with Mr Trump. Privately, officials fret about what would happen if America—bent on achieving peace at any cost—were to forbid allies from sending American-made weapons and parts to Ukraine, cut off vital battlefield intelligence or threaten to pull its troops out of the Baltics or Poland. Politicians in Stockholm express hopes that America knows it would pay too high a political price. A senior figure suggests that Mr Trump may now have lost interest in Ukraine, preferring to focus on Iran and China.

Ukraine’s allies cannot easily see a path to the war ending well. They fear a bad peace, for Ukraine and for their own security. They distrust Russia absolutely, and expect to do so for years to come. A stalemate, then: a bad outcome better than all others. ■

ch103.xhtml

How Jensen Huang became AI’s Taylor Swift

When the Nvidia boss returns to his native Taiwan, things get a little weird

Sept 3rd 2026

For a few days in June, Taipei, the Taiwanese capital, succumbs to a kind of fever. Hundreds of people throng outside restaurants in the sweltering heat after hearing rumours that Jensen Huang will be dining there. Every few minutes a passing black car prompts a flutter of excitement. Perhaps he is inside.

The boss of Nvidia, the world’s most valuable company, comes back to his native Taiwan every summer for Computex, Asia’s biggest computer-hardware conference. During his stay he will typically peel off to eat at street stalls, walk around and chat to people. One evening this year he moved amiably through the Raohe night market, sampling matcha desserts and joking with the owners of a grilled-corn stand that he’d buy dinner for everyone in the queue if they’d let him skip to the front. Crowds of Taiwanese followed in his wake, hoping for a selfie, an autograph or an exchange of banter. Later that night Huang wrote “Jensen was here” on the tiles of a restaurant toilet. Straight away a queue formed: people wanted to photograph the precious piece of Jensenalia.

On another evening Huang went to a restaurant with the bosses of some South Korean technology companies. As he stepped out of the car, a crowd surged forward, phones aloft; his security detail struggled to contain the crush. The technology executives looked bewildered, as though they had accidentally wandered into the wrong event. Huang, by contrast, seemed entirely at ease. He smiled, waved and posed for the cameras before disappearing inside. “His skin is so smooth,” a fan next to me sighed. Some people in the crowd had T-shirts and key rings with Huang’s face on them, which were doing a roaring trade at the convention centre.

“His skin is so smooth,” a fan next to me sighed

In the AI gold rush, Huang is the man with the shovels. Nvidia, the company he co-founded, designs and sells the ultra-powerful chips which provide about two-thirds of the world’s AI computing power. Bosses of the world’s biggest technology firms and governments racing to build their own domestic AI infrastructure court him to get their hands on Nvidia’s coveted processors. The markets are in thrall to him: a passing compliment from the leather-jacketed Nvidia boss can send a company’s shares soaring.

Yet that alone does not explain the mania Huang excites in Taipei. Mark Zuckerberg and Sam Altman are powerful technology bosses, but it’s hard to imagine Americans queuing for a glimpse of either, or buying T-shirts featuring their faces (Zuckerberg once described Huang as “like Taylor Swift, but for tech”).

Jensanity, as the fervour is known in Taiwan, is partly local-hero worship. Huang, who was born on the island before emigrating to America, is Taiwan’s greatest business success story. He has also done matchless international PR for the island’s electronics industry, calling it the “epicentre” of the AI revolution. Most of the world’s advanced computer chips are made by one Taiwanese firm, the Taiwan Semiconductor Manufacturing Company (TSMC). Politicians on the island, which only a handful of governments recognise as a sovereign country, hope its critical role in the AI boom will give it a measure of soft power. This is sorely needed: increasingly assertive behaviour from China means that Taiwan’s security rests, to a large extent, on the willingness of others to defend it—especially America.

Huang hasn’t just raised Taiwan’s profile; he has also made many Taiwanese a lot richer. “Jensen has been great for my portfolio,” one young engineer waiting for a glimpse of Huang outside a restaurant told me. The TAIEX—the island’s benchmark stock index—has risen by 90% over the past year, making it the fifth-largest stock market in the world. Much of that increase reflects the extraordinary rise of TSMC, which manufactures chips for Nvidia. Since 2023 the Taiwanese firm’s market value has increased more than fivefold, to around $2trn, and dozens of its local suppliers have prospered alongside it.

But, as I observed over three days shadowing Huang, Jensanity is also stoked by the Nvidia boss himself. Unlike other Silicon Valley heavyweights, Huang goes out of his way to court the public. One evening in Taipei, while dining in a restaurant with his parents, he emerged carrying trays of steamed buns for the journalists waiting outside.

The public persona is useful to his business. Analysts told me Huang’s rockstar status makes Taiwanese manufacturers eager to work with him, hoping his stardust will rub off on them. That gives Nvidia an advantage when AI supply chains are tight and chip companies are chasing manufacturers, rather than the other way round. After this year’s Computex Huang flew straight to Seoul, where he wooed the bosses of South Korean memory-chip companies over beer and fried chicken, and was filmed by a local TV channel dancing—a little awkwardly—to K-pop. Not for nothing have the Taiwanese given Huang the affectionate nickname, “the People’s Dad”.

When I first attended hardware conferences like Computex nearly two decades ago, as a chip designer, they were dull affairs. Socially awkward engineers gathered to inspect server racks, debate thermal specifications and negotiate component prices. Journalists were interested in Microsoft and Google, and the Taiwanese chipmakers that powered the big tech companies’ gadgets remained largely invisible outside the industry.

So far, Huang has kept the intricate supply chain humming. But other challenges are mounting

But training frontier AI models requires a staggering amount of computing infrastructure, and in the past few years hardware-makers have been thrust into the limelight. The buzz at Computex was palpable, not just owing to the drones flying through the conference centre. Instead of the desultory exhibition stands I remembered, there were giant, neon-lit pavilions. Humanoid robots wandered the aisles, as exhibitors handed out tote bags and bubble tea. Computex now has a headline act. The queue to get into the main conference hall for Huang’s speech began hours beforehand, and stretched thousands-deep.

When he finally appeared on stage in his trademark black-leather jacket, the hall erupted. Huang began by thanking his suppliers and partners in Taiwan, and logos of the usual electronics manufacturers flashed up on the screen behind him, suggesting they were going to get a shout-out. The first “supplier” he name-checked, however, was a fruit seller at the Tonghua night market. A local dumpling shop came next, followed by a succession of other eateries. The audience lapped it up.

Huang’s journey to the centre-stage of the global economy was an unconventional one. Born in Taiwan in 1963, he was sent to America at the age of nine together with his older brother to live with their uncle. The uncle, mistaking a reform school in Kentucky for an elite boarding school, enrolled them there by accident. Huang slept in a dormitory and was expected to clean its toilets, but was too young to go to any classes there. Instead he studied at the local school, where the other children bullied him for his accent and called him a “chink”. Huang has credited those years with toughening him up.

Two years after the boys arrived in America, their parents flew out to join them and the family settled in Oregon, where Huang studied electrical engineering before heading south to Silicon Valley. He founded Nvidia in 1993 with two fellow engineers. The chips that would make Nvidia famous, graphics processing units (GPUs), were then a product used mainly for video games, and the market was crowded. Nvidia became dominant by designing better chips and, crucially, getting them to market first.

By the mid-2010s, Huang had become convinced that AI would transform computing and that GPUs, whose ability to perform many calculations simultaneously made them well suited to AI models, would become its essential building blocks. For years that was an unorthodox position to take. AI was always touted as the next big thing, but its moment never quite seemed to arrive. In November 2022, OpenAI, a modelmaker, released ChatGPT. Demand for Nvidia’s chips, on which ChatGPT was trained, exploded. Nvidia’s market value climbed from $340bn to $5trn in just three years.

Steve Bannon has called Huang an “agent of influence” for the Chinese Communist Party

The company’s success rests on more than clever chip design: it’s also dependent on the ability to manage relationships with supply-chain partners, and the political environment which affects them all. Heroes of software companies are often portrayed as lone geniuses, but hardware-making is an inherently collaborative project. Nvidia’s latest product, Vera Rubin NVL72, is a complete AI supercomputer for training and running large AI models. It contains 1.3m components, more than 1,000 chips and the software needed to harness them. Building one requires hundreds of suppliers, manufacturers and software companies to work together.

The AI boom has led to a scramble for those components, and a big part of Huang’s job is ensuring that suppliers prioritise Nvidia’s orders. Money helps. Nvidia has used its enormous balance-sheet to sign multi-year deals with manufacturers. Huang also needs his suppliers to make more. Analysts believe Huang has pushed back several times against TSMC’s caution about investing in new capacity.

So far, Huang has kept the intricate supply chain humming. But other challenges are mounting, not least the threat of competition from Nvidia’s biggest customers. Alphabet, Amazon, Meta and Microsoft are each seeking to reduce their dependence on Nvidia by developing their own AI chips. Nvidia’s stock price meanwhile has become a proxy for investor sentiment about how long the AI boom can last, and has dipped several times this year. Huang’s answer is to bind Nvidia even more tightly to the future of AI. Rather than selling only chips, he also wants to provide the infrastructure, software and AI models on which the industry runs.

This strategy will require even more supply-chain diplomacy, and I wasn’t surprised to see Huang schmoozing manufacturing partners in the exhibition hall during his time at the conference. His staff explained that he made a point of visiting suppliers’ booths because his presence draws crowds and gives them publicity. Shouts of “Jensen! Jensen!” echoed through the hall as he made his way from stand to stand to chat to engineers and executives, high-fiving people along the way. Only late in the afternoon did he seem to tire, sitting cross-legged on the floor of a booth to share a beer with a Taiwanese executive. Hundreds of fans crowded round to capture the moment on their phones.

After his keynote speech I sat down with Huang in a private room in the conference hall. By then he had shed his leather jacket for a navy cotton one. Outside, his staff and security were bustling about, but Huang seemed calm and relaxed.

When I explained that I wanted to understand him rather than Nvidia, he looked faintly embarrassed. He was uncomfortable talking about himself, he said, and apologised in advance if he did a poor job of answering my questions. On stage he’s an exuberant performer, but in private conversation he almost whispers, speaking in slow, measured sentences.

I asked how his life had changed since the launch of ChatGPT. Silicon Valley lore is full of founders who become celebrities overnight. Huang’s rise has been about three decades in the making, which may have made the transition to superstardom even stranger when it finally came. Nothing, he insisted, had really changed. He dressed the same way and, as he put it, still “puts his pants on one leg at a time”.

Analysts told me Huang’s rock-star status makes Taiwanese manufacturers eager to work with him

Huang hasn’t just had to get used to being famous; he has also found himself at the centre of a geopolitical contest between America and China. Since October 2022 the American government has been trying to stop its adversary from obtaining advanced computer chips. Nvidia’s suppliers in Taiwan, meanwhile, live under the growing threat of military action by China. Huang suddenly found he needed to build rapport with world leaders, not least Donald Trump, whom he has lobbied to relax restrictions on chip exports.

Hanging out with politicians doesn’t come naturally to him, he said. He has nonetheless proved adept at translating his goals into language that appeals to them. He recasts data centres as “AI factories”—a term that appeals to Trump’s enthusiasm for manufacturing. To foreign governments wary of relying on American companies, he has offered “sovereign AI”—national infrastructure built around local values, and largely powered by Nvidia.

In May he demonstrated his knack for diplomacy when he joined Trump on a trip to China for a summit with President Xi Jinping. Huang appears to have been a last-minute addition to the entourage—many were surprised when Trump announced on Truth Social that “the Great Jensen Huang” was on board Air Force One. In Beijing Huang met Chinese officials and accepted a seat on the advisory board of Tsinghua University, one of China’s most prestigious institutions.

China hawks are suspicious. Steve Bannon, a former White House chief strategist, called Huang an “agent of influence” for the Chinese Communist Party and demanded his arrest. Elizabeth Warren, a Democratic senator from Massachusetts, has said that Huang’s lobbying for the sale of advanced AI chips to China could “turbocharge China’s military and undercut American technological leadership”.

Yet, oddly, for a supposed China-appeaser, Huang has acted as a public booster for Taiwan. He says Taiwan made Nvidia possible and calls it his “business home”. Nvidia is building one of its largest offices outside America in Taipei and plans to invest around $150bn every year. The balancing act doesn’t always go smoothly—in 2024 Huang angered China by describing Taiwan as “one of the most important countries in the world” (he was forced to clarify that he had been praising the island’s electronics industry rather than recognising Taiwan as a state). Being an AI hardware titan in today’s world, he told me wryly at Computex, is “the biggest communications job ever”.

Huang doesn’t rely on charm alone to get suppliers to do what he wants. According to Rick Tsai, a former boss of TSMC, who has worked with Huang for decades, he is also “hard driving” and “relentless”. Tsai calls Huang’s approach to the supplier relationship “rough justice”: an understanding that sometimes one side is right, sometimes they are wrong, but over the long term, the wins and losses roughly balance out.

Huang hasn’t just raised Taiwan’s profile; he has also made many Taiwanese richer

This same brand of rough justice appears to govern life inside Nvidia. Employees say Huang has little patience for excuses and is known to challenge people when something is not working. Around 60 people report directly to him, an unusually high number for a chief executive. He gives feedback immediately and publicly, reasoning that the entire room can learn from the bruising.

I occasionally saw traces of the unfiltered style. At a press conference, Huang rebuked a staffer who moved too quickly to snatch a microphone from an analyst. When a BBC reporter asked about the geopolitical risks facing Taiwan, Huang initially offered a well-rehearsed answer. When the reporter continued to press him he joked that someone should help the BBC come up with a new question. The audience laughed, but Huang’s annoyance was evident.

Rev Lebaredian, an engineer who has worked at Nvidia for more than two decades and is now a vice-president there, said that Huang’s executive team meetings would look “miserable” to an outsider. All the talk is about problems and things that are failing; good news is rarely discussed.

Yet according to Lebaredian, Huang has created a culture where employees know that although their mistakes will be called out, they will not be punished for making them. Nvidia’s senior leadership team has remained stable over the past two decades, a rarity in the industry. They have also been well rewarded. Huang likes to boast that he has created more billionaire executives than “any CEO in the world”. Rank-and-file employees have not done badly, either.

When I asked Huang what drove him, he didn’t speak about colonising other planets, defeating death or remaking humanity. Elon Musk was a friend, he noted wryly, but they had different goals. His, he said, were simpler: he wanted his employees and partners to do well. “Maybe,” he said with a smile, “I’m just not sufficiently ambitious.” ■

Shailesh Chitnis is a global business writer at The Economist.

Photographs by Leong Chong Lao

ch104.xhtml

Will anybody use AI as much as coders do?

The answer will have big implications for the investment boom

Illustration: Simon Bailly

Aug 30th 2026

How software engineers at Silicon Valley startups spend their days has changed dramatically over the past few years. The output is still the same: reams of code. But instead of sitting and typing, they whisper into microphones. Their instructions are then transcribed using artificial intelligence and fed to AI agents—bots that can interact with software and, in this case, write it.

Other white-collar workers are not yet “in this world”, admits Colin Jarvis of OpenAI, maker of ChatGPT. The future of his employer—and the AI boom at large—hinges on that changing. To justify their colossal outlays on data centres, America’s cloud giants, and AI labs such as Anthropic and OpenAI, need adoption of the technology to continue spreading rapidly. A back-of-the-envelope calculation by The Economist finds that covering the cost of the data-centre build-out will require total annual revenue from the use of AI to reach $2.5trn per year by the end of the decade, up from around $150bn now.

So far uptake has been strongest by far among software developers. Four-fifths of programmers say they use an AI coding tool, according to Stack Overflow, an online forum. SemiAnalysis, a research firm, reckons that for Anthropic and OpenAI coding now accounts for over half their combined annual recurring revenue—a trendy measure of sales that extrapolates from the most recent month. In June 2025 the combined annual recurring revenue of Cognition, Cursor, Lovable and Replit, four AI coding startups, was roughly $800m. Today it stands at $6bn.

Chart: The Economist

Coders’ enthusiasm explains why AI has diffused more quickly in industries where they account for a larger share of employment (see chart). Optimists may see this as evidence that the AI boom is just getting started, with spending on the technology reaching far greater heights in the years ahead as other applications mature. Yet there may also be structural reasons why coding is an outlier.

Three candidates often touted as the next runaway application for AI are lawyering, finance and customer service. Growth in all has been impressive. Harvey, Clio and Legora, three developers of AI legal tools, have roughly doubled their combined annual recurring revenue in the past year to $1bn. Harvey says the amount of time lawyers spend on its platform is doubling each month. Bankers and other finance types are also starting to embrace the technology. Rogo, an AI startup building tools for the financial-services industry, added 100 enterprise customers in the past quarter and increased annual recurring revenue by 50%.

AI is also taking off in customer service. Sierra, which builds tools for the purpose, reached an annual recurring revenue of $200m in June, double the figure in November last year. Venture capitalists are bullish. So far this year they have poured $3bn into AI startups in this area, more than for any other category of AI application. (Investment in AI coding startups has stalled, in part because AI labs’ own coding tools have grown so quickly.)

These three applications bear some similarities to coding. They are big potential markets, and the work is often formulaic. But four factors set coding apart: the availability of training data; how easy it is to test a model’s output; the low amount of human interaction involved in the work; and software engineers themselves. AI companies are trying to overcome the differences in other markets, but doing so will not be straightforward.

Part of why AI has proved so successful as a programming buddy is that a huge amount of open-source code has been published on the internet. It has thus made its way into models, accounting for almost a fifth of all training data in some cases. The same volume of publicly available content simply does not exist for customer complaints or contract negotiations.

Code generated by AI is also relatively easy to validate. Software tests are a routine part of engineering: one piece of code checks another. That increases the likelihood that any errors in AI-generated code are caught before they wreak havoc. Moreover, testing helps to improve the performance of models by alerting them when a mistake is made. Assessing whether an irate customer was appeased or a valuation model appropriately reflected the nuance in a company’s accounts is trickier.

Software-writing tools have the added advantage that much of the information required to complete a task already exists in a program’s code-base, or can be reached through the digital pipelines that connect software, known as application programming interfaces. In many other contexts crucial information is not available in digital form, instead residing in the brain of a call-centre worker or the facial expression of a counterparty’s lawyer.

AI companies are working on all these challenges. Winston Weinberg, the boss of Harvey, says that advances in AI mean his firm is now able to train its tools in part by relying on “synthetic” data, which are generated by AI systems. Startups such as Mercor are connecting AI companies with professionals in various fields who can transfer their expertise into models. Ping Wu, boss of Cresta, an AI customer-service startup, says that by analysing thousands of call-centre transcripts his firm can infer some of the knowledge lodged in the heads of staff, such as how refunds are handled. Legions of so-called forward-deployed engineers are now being dispatched by AI companies to their customers’ offices to troubleshoot problems and improve their tools.

But the last factor behind AI’s success in coding may be harder to replicate: software engineers themselves. Being both nerdy and accustomed to change, given that programming languages are updated every few years, they are natural adopters of new technology. Much of the uptake of AI coding has thus happened in a bottom-up way, with engineers discovering and experimenting with the tools of their own accord, notes Anish Acharya of Andreessen Horowitz, a venture-capital firm.

Call-centre agents, by contrast, typically have neither the incentive nor the freedom to experiment on the job, and tend to wait for technologies to be deployed top-down. Lawyers and finance workers, for their part, must tread carefully to avoid falling foul of the many regulations that govern their occupations.

Other applications of AI will certainly continue to grow, and may well accelerate as models improve and companies become more comfortable with handing over tasks to AI agents. But it could be a while before the majority of office drones spend their days whispering into microphones. ■

ch105.xhtml

Neoclouds like CoreWeave are getting much bigger—and riskier

The mountain of debt funding their expansion is growing rapidly

Photograph: Getty Images

Sep 3rd 2026

New technologies often create good conditions for middlemen. In the 1960s entrepreneurs snapped up mainframes from IBM and leased them to banks and defence companies. In the dotcom era “domainers” bought website addresses and sold them at a higher price. Artificial intelligence is no different. Consider the neoclouds, such as CoreWeave, which buy specialist AI chips and rent them out. They are among the fastest-growing companies in the AI industry. But they are also some of the most vulnerable.

Neoclouds have thrived over the past few years, for two reasons. The first is that demand for AI has grown more quickly than America’s “hyperscalers” have been able to expand their computing capacity. Many neoclouds started off as crypto miners, and subsequently repurposed their data centres. And because they have focused exclusively on AI, rather than a broader range of cloud services, they have been able to concentrate their investments. As a result, their biggest customers are model-makers, such as OpenAI, and the hyperscalers themselves, who often lease extra capacity from the neoclouds. The second reason for their rise is that Nvidia, the leading maker of AI chips, which is eager to reduce its reliance on the hyperscalers’ custom, has provided various forms of financial support to help them buy more of its chips.

The combined annual revenue of the five largest listed neoclouds—CoreWeave, Nebius, IREN, Applied Digital and Core Scientific—has risen to $18bn over the past 12 months, up from $3bn in 2024. Although they are still far behind the hyperscalers, they are gaining ground. Dell Oro, a research firm, estimates that in total neoclouds have 1.5-2.5 gigawatts (GW) of data-centre capacity between them, compared with a combined 20-24GW for Alphabet, Amazon, Microsoft and Meta. But they are roughly doubling their capacity each year, which is about triple the rate that the hyperscalers are managing.

Chart: The Economist

That growth, however, is being funded by mountains of debt. The five listed neoclouds now have $61bn-worth of borrowing on their balance-sheets, quadruple the amount just 12 months ago, as well as $18bn-worth of operating leases (see chart). None is generating an operating profit, and all are deeply in the red once interest bills are added. In the most recent quarter Nebius’s interest payments amounted to a fifth of its revenue; CoreWeave’s reached a quarter.

The neoclouds are betting that, if they continue their vertiginous growth, profits will eventually arrive. Perhaps. But the competition has shown no sign of easing. Besides the established cloud giants—as well as Meta, which is considering launching its own third-party service—there are now some 200 neoclouds around the world duking it out for contracts.

Add to that a growing cohort of mid-sized providers. Oracle, originally a database company, has spent a fortune trying to join the hyperscalers’ club with a cloud business of its own. Its capital expenditure, much of which is financed by borrowing, is expected to hit $77bn this year. In July SoftBank, a Japanese conglomerate up to its eyeballs in debt, unveiled a cloud business called SB Neo. The group also owns a majority stake in SB Energy, an infrastructure business that has teamed up with OpenAI to build a giant data centre in Ohio, and which filed on September 1st for an initial public offering.

Fierce rivalry compounds the two big risks facing neoclouds. The first is a credit-market shock that pushes borrowing costs higher, further squeezing margins. Investors already seem jittery: the spread in yields between neoclouds’ bonds and junk bonds as a whole is widening.

A second—and even bigger—danger is that supply and demand rebalance. Right now, vast amounts of infrastructure are being built on the assumption that use of the technology will continue to soar. If the pace of adoption slows, however, model-makers and hyperscalers alike may quickly decide they no longer need the neoclouds.

Either event would be cataclysmic for the industry. Only those neoclouds that offer something more to customers than sheer capacity would avoid going bust. To that end, some are focusing on positioning themselves as reliable infrastructure providers for Western governments that have grown suspicious of America’s tech giants. Others, including CoreWeave and Nebius, are trying to offer added software services to customers. To survive, the neoclouds may need to be more than middlemen. ■

ch106.xhtml

From protein to matcha, food fads spread faster than ever

Supply chains struggle to keep up

Demand whey up © Getty Images

Sep 3rd 2026

PIGS WERE once the biggest consumers of whey, a byproduct of cheesemaking. Now farmers can scarcely afford it. Demand for whey concentrate, a common ingredient in gym shakes, has soared amid the craze for consuming more protein. Dairies have not scaled up cheese production to match. The price of whey protein thus rose fivefold in the three years to June, according to Vesper, a research firm. Cottage cheese, which TikTok influencers also recommend for bulking up, has seen a similar shortage.

Both cases illustrate a new problem for the food-and-drinks industry. Online trends can build demand for ingredients faster than ever before, but farms and factories cannot scale production at the same speed. Firms are adapting by learning to anticipate trends and ride them out before consumers move on.

Food fads are hardly new. Pineapples were so coveted in Georgian England that rental shops opened peddling display fruits. Swiss fondue, a hit with Americans in the 1960s and 70s, resulted in millions of cheese-melting pots that now languish in cupboards. Today, though, social media spreads such trends across the world in weeks. When a video of a pistachio-pastry chocolate bar made in Dubai went viral on TikTok in 2023, it was viewed more than 100m times. Months later top chocolatiers were selling their own versions, straining global pistachio production.

Food and drinks are especially prone to going viral. Grass-green matcha or violet ube lattes make for pleasing Instagram snaps. They are also cheap enough to be accessible. Bought once, even a $19 strawberry from Japan (another recent hit) won’t break the bank. Young buyers, who spend a good deal of their lives on social media, are also more inclined towards novelty, notes Dan Frommer, author of “The New Consumer”, a blog.

Explosive demand has overwhelmed supply chains. Tencha leaves, which are used to make matcha, nearly tripled in price between 2024 and 2025 amid surging interest in the drink. The bushes from which they come require years to grow. Hot markets also attract chancers. Rocky Xu, a matcha seller in Los Angeles, says that when he ran out last year, he found that the market had become crowded with middlemen. The newcomers—mainly Japanese entrepreneurs hoping to mediate between farmers and foreign buyers—pushed up prices.

Food companies are trying to cope in several ways. Anticipating trends as early as possible allows them to lock in suppliers before rivals do. Starbucks, an American coffee chain, started offering green-tea lattes in 2006, long before the matcha boom in the West. It popularised drinks with “popping pearls” in 2024, before bubble tea went mainstream. More recently it added a protein range to its coffees, which are now as popular as flat whites in Europe. Louise Direito, an executive at the chain, says that it has switched from relying on “small focus groups” to tracking “millions of customers shaping trends in real time” on the internet. “The challenge is knowing when to join the conversation, as well as when it’s too late,” she adds.

Collaborations also help. As spicy foods became all the rage in 2022, fast-food chains lined up to work with Fly by Jing, a trendy maker of chilli-oil. Shake Shack, a burger chain, built a whole chicken menu around it.

But not all fads are worth embracing. Starbucks’ olive-oil coffees were pulled from menus in America and Canada less than a year after their launch in 2024—amid complaints about laxative effects. ■

ch107.xhtml

Chinese businesses of all stripes are turning to Huawei

In industries from mining to transport, the company’s technology is becoming pervasive

Friends of mine © Getty Images

Sep 3rd 2026 | Hailar

Driverless taxis are a common sight in Chinese cities these days. So, increasingly, are delivery drones. But to see one of the heaviest autonomous machines ever deployed, head to the grasslands in China’s far north-east, to the Yimin coal mine, a giant hole in the earth spanning 42 square kilometres. Travel to the bottom and you will find around 100 lorries each capable of carrying 90 tonnes of coal—without a human in sight. Large black boxes filled with batteries sit where the driver’s cabins should be.

Zhao Yaozhong, a manager at Huaneng, the state-owned power utility that operates Yimin, says the deployment of driverless lorries has led to 120 workers being reassigned to posts outside the pit—thus removing them from the dust and danger of the mine. The vehicles also do far more work without a human driver. They run through the night and in poor weather conditions, and even change their own batteries. If all goes to plan, most humans will disappear from the pit over the next three years as another 300 autonomous lorries are added.

The operation at Yimin offers a glimpse into the expanding “enterprise” division of Huawei, a Chinese technology giant better known for its telecoms equipment and consumer electronics. In 2021 Huawei, whose core business had been hit hard by American sanctions, announced that this division would create a collection of new teams—referred to as juntuan (Chinese for army corps)—that would focus on applying expertise and technology from across the business to specific groups of enterprise customers, from airport operators to oil-and-gas companies.

The mining corps shows this integrated approach in action. Huawei’s telecoms base stations, which are visible throughout Yimin, cover the area with 5G, allowing the lorries’ autonomous-driving system to function. That system was based on one developed by Huawei’s automotive division, but was tailored to suit the size and weight of the lorries and the conditions within a mine. The interface used to control the mine’s operations runs on an operating system called MineHarmonyOS, which was developed by the same division that makes Huawei’s smartphones and other consumer gadgets. In a number of underground mines, an artificial-intelligence model created by Huawei’s cloud division is used for interpreting geological data and remotely monitoring equipment.

Huawei, which is privately held, has not made it easy to understand how its enterprise division has fared. Revenue was last broken out in 2022, at which point it accounted for a fifth of total sales and was the fastest-growing part of the group. Andy Wu, who runs the mining unit, says his business is profitable, but that margins are extremely thin.

Nevertheless, the enterprise division is set to play an increasingly vital role for the company. On August 31st Huawei said that its net profits fell by 37% year on year in the first half of 2026, even as revenue grew by 10%. Margins in its consumer division have been squeezed by the surge in memory-chip prices resulting from the data-centre boom. Huawei has also been investing heavily in research and development—which gobbled up a quarter of revenue in the first six months of the year—as it looks to replace American technology it has lost access to, including the most advanced semiconductors.

Concern that Huawei’s products could be used by China for snooping has led many Western governments to impose restrictions, causing the company’s foreign sales to shrink over the past decade. That has made it all the more important to find new avenues for growth at home, particularly now that the build-out of the country’s 5G network has largely concluded. Fortunately for Huawei, China’s government is enthusiastically championing real-world applications of AI in domains from manufacturing and transport to the operation of electricity grids. Huawei’s enterprise corps stand ready. ■

ch108.xhtml

The IRS is going after America Inc’s overseas profits

Companies from Coca-Cola to Meta are in its sights

Photograph: Getty Images

Sep 3rd 2026 | New York

For decades American businesses have fattened their bottom lines by shifting profits to low-tax jurisdictions abroad. In 2025 alone they saved at least $11bn using these manoeuvres, reckons the FACT Coalition, an advocacy group. Now the government wants some of that back.

Many American bosses breathed a sigh of relief when, on his first day back in the Oval Office last year, Donald Trump announced that the country was abandoning a global agreement orchestrated by his predecessor that would have subjected tax-shy multinationals to top-up duties. But those who hoped this would mark the end of government efforts to reclaim a greater share of their profits booked abroad were mistaken. In May this year UnitedHealth, an American insurance giant, disclosed that the Internal Revenue Service (IRS) was pursuing it over profits it had attributed to a foreign subsidiary, with the agency insisting it had underpaid its taxes back home. The case is part of an effort by the IRS that began during the Obama administration to crack down on so-called transfer pricing—an initiative the Trump administration has shown no sign of resiling from. Many billions of dollars are at stake.

Among the companies the IRS is tussling with over transfer pricing are Amgen, a biotech firm, and Meta, owner of Facebook and Instagram. But it is the case against Coca-Cola, the world’s most valuable drinks company, that has been called the “Super Bowl of transfer-pricing controversy”. (Tax lawyers are easily entertained.)

That case, which has been in train for over a decade, may at last reach a conclusion in the coming months. Back in 2015 the IRS told Coca-Cola it owed an additional $3.3bn in tax for the period from 2007 to 2009, arguing that the company had improperly parked profits arising from intellectual property—including its famous logo—overseas. The battle has taken years to wind its way through the courts, during which time Coca-Cola has continued to follow the same tax practices. As a result, the money at stake has ballooned to as much as $20bn. Coca-Cola argues that its foreign subsidiaries have made valuable marketing investments worthy of the astronomical returns on capital it has attributed to them. It also contends that it is merely sticking with an agreement the company reached with the IRS back in 1996 on how the company could split its profits (though that deal related to past earnings).

Coca-Cola and others pursued by the IRS over transfer pricing have been remarkably relaxed about the situation, at least publicly. As of July the drinks giant had squirrelled away just $530m to pay a tax bill that could amount to as much as $14bn in the event of a loss (the company has already shelled out $6bn, which it would be refunded if it wins). In July Amgen settled an investor lawsuit that accused it of underselling the risks of its fight with the IRS. Its disclosures were like “a child telling his parents that he had ‘dessert’ when in fact he had eaten the ‘whole cake’”, noted a federal judge during the case.

Winning its largest transfer-pricing cases would allow the IRS to claw back roughly $100bn in additional taxes from a small group of firms, including penalties and interest, notes Steven Wrappe of Grant Thornton, an accountancy. That amounts to about a fifth of the total corporate-income tax that the American government collected in 2025.

In the short term, expanding the crusade would be difficult; the IRS’s budget has been slashed under Mr Trump, meaning its staff are already stretched. In the long term, however, transfer-pricing disputes will probably increase. One reason is the growing share of firms’ value that comes from the kind of intangible assets that facilitate cross-border tax dodges, because they do not need to be physically moved. From 1975 to 2025 the share of the total value of firms in America’s S&P 500 index that can be attributed to intangible assets rose from 17% to 92%, according to Ocean Tomo, an advisory firm. Another reason to expect more of these cases is that—as bond-market investors know too well—America’s government debt continues to soar. Its multinationals may soon have to foot a bigger share of the bill. ■

ch109.xhtml

LIV Golf’s failure shows how not to invest in sport

It takes more than huge amounts of money to disrupt a game

Photograph: Getty Images

Sep 3rd 2026 | LOS ANGELES

PUTTING ASIDE the exorbitant salaries, the silliness of the “Golf but louder” slogan and the din of pop music on the fairway, there was always something a bit plucky about LIV Golf. Saudi Arabia, a country that looks like a bunker but has no golfing tradition, was attempting to disrupt one of the world’s most snooty sports. Predictably, its attempt ended in failure. Last month the Public Investment Fund (PIF), Saudi Arabia’s $900bn sovereign-wealth fund, pulled the financial plug. In the coming days LIV Golf is expected to file for bankruptcy.

It is not just golf, though. The PIF has led a wave of investment by the kingdom over the past five years in football, tennis, boxing, martial arts and motor racing, plus e-sports (as video-game tournaments are known). Its ambition was impressive. Many accused Muhammad bin Salman, the Saudi crown prince, of “sportswashing” his country’s thuggish reputation by pouring petrodollars into athletes’ pockets. They missed the point, at least partly. The PIF hoped to use the thrill of sport to make young Saudis more active, as well as hosting marquee events to attract tourists and diversify its oil-soaked economy.

Now, however, physical sports are in the relegation zone. The PIF’s recently published strategy for 2026-30 mentions the word “sports” only seven times in 50-odd pages—on six occasions prefixed by “e-”. At a time when private-equity firms and other investors are pouring so much money into sporting franchises that some perceive a bubble, what can be learned from the Saudis’ retreat?

The first lesson is the importance of fans. It is one thing to have a near-bottomless bank account. It is another to buy your way into supporters’ hearts. The PIF blew $5bn on LIV Golf hoping to create a rival to America’s PGA Tour, using lavish prize money, faster games and more razzmatazz to attract younger fans. But though it signed some star players, LIV turned off golf’s well-heeled supporters, who like the PGA’s genteel traditions and prestige. With too few fans, it could not generate the broadcast income to be sustainable.

Fans can also be powerful foes. FIFA, football’s global governing body, recently scrapped an attempt to raise $4.2bn by selling off chunks of the beautiful game, including the World Cup. That prompted a backlash among national federations, which feared, among other things, football supporters’ wrath.

Therein lies the second lesson. In sport, creative destruction is hard. As with many enterprises that have been around for generations, it is tempting to think that sports are ripe for disruption. Besides the PIF, others have sought to shake up established leagues by launching snazzy alternatives, only to fail. Grand Slam Track, an alternative to the Diamond League in athletics, was started by Michael Johnson, an American sprinter. It filed for bankruptcy in December less than a year after its first fixture. Only a few upstarts prosper, such as Ultimate Fighting Championship, an American mixed-martial-arts organisation that has taken cage fighting to a new level. It created a new market rather than trying to muscle into already popular terrain.

The third lesson is time—or timing. Building a winning sports franchise can take generations. More than 50 years have passed since Pelé, a great Brazilian footballer, first played for the New York Cosmos. Only now is American soccer coming into its own (and Pelé’s Cosmos have long been defunct). The quickest way to make a buck today may be to buy a well-known franchise and sell it shortly afterwards. That is what Mark Walter, a financier, did with the Los Angeles Lakers, selling the basketball team last month to two fellow billionaires for $12.5bn, $2.5bn more than he had paid for it a year earlier. But in a frothy market, flipping is a risky strategy.

If all else fails, think ahead. Saudi Arabia has a mixed record investing in sports that have been around for centuries. But in the past five years it has taken a more methodical approach to the emerging field of e-sports, including leading a $55bn buy-out of Electronic Arts, a gaming powerhouse, which was completed last month. Watching gaming tournaments may not get young Saudis off the couch—but they love it. As an investment, it is less quixotic than LIV Golf. ■

ch110.xhtml

The glory of the corporate training video

Cartoons. Gold stars. Terrible acting. It’s all there

Photograph: Paul Blow

Sep 3rd 2026

Workplace training can confer useful skills. It can also prevent legal trouble and fulfil compliance obligations. And it can be a very easy way to try out time travel. You only have to watch a typical corporate training video to feel the clock rewinding.

First, and most obviously, some of these videos were filmed before you were born, on a subject that is so evergreen it has never needed an update. The footage is grainy. Men wear ties and women wear pleated skirts. They have names like Rodney and Carol. No one has a tattoo or brings their whole self to work. If they had an illness, it wouldn’t be a mental-health issue but something from the distant past, like consumption or scrofula.

Some training videos may have been made recently but still look incredibly dated. That’s because they are produced on the cheap, feature terrible actors who cannot find other work and have been given the remit of achieving the longest possible shelf-life by not showing anything too contemporary.

Watch a video on harassment at work, for example, and you are likely to see men in high-viz jackets pretending to shout at each other while a narrator warns that “horseplay” can get out of hand. Horseplay? Harassment can leave people feeling isolated, you are told, as someone sits in a digger and looks isolated. You haven’t seen performances this bad since a nativity play.

Second, plenty of training materials are on subjects that require no training. To watch some corporate videos is to be treated like a particularly dim child—or, more accurately, like someone a particularly dim child would consider to be a moron. Your columnist recently completed an online video course on workplace slips, trips and falls, and was left wondering how the species has managed to survive so long.

There were warnings about wet leaves and the dangers of transitioning from outdoors to indoors. There were moments set aside for deep reflection on questions such as: “What are the most common things that you slip or trip on in your typical day-to-day [sic]?” And there were remarkable new insights. It turns out that slips can be avoided by ensuring you wipe the soles of your shoes on a mat when entering the premises of a building. Note that word “soles”. You can spot an untrained worker a mile off by the way they lie down and flap around like an elephant seal on heat as they try to dry the tops of their shoes.

Third, lots of training videos lean very heavily on the techniques of primary schools. When you’re not watching wooden actors, you’re being fed cartoons. Every so often, multiple-choice questions test your level of understanding. The slips, trips and falls course features a waiter who had not noticed that “the decorative rug at the entrance to the dining area had lost its grip.” He had caught his foot on the edge of the rug, stumbled and regained his balance. What type of incident had he experienced: a slip, trip or fall?

It’s hard to conceive of a lower-stakes scenario than a waiter not falling over, but your inner toddler is given an incentive to care. If you get a multiple-choice question right on a training video, you will probably win a badge or a gold star. Ace the entire test, and confetti will fill the screen and you will get a downloadable certificate of achievement. Fail to finish the course, and you will be reported to the teachers.

Treat an audience as if they are children and their behaviour will revert to childish norms. You might be approaching retirement but you still get a tiny thrill from getting a gold star. You become insanely proud of your slips, trips and falls certificate, and pin it to your cubicle wall. You wonder to yourself whether you could take the course on how to use stepladders, and shake your head at your own wild ambition.

You also succumb to the urge to cheat. You start playing a cyber-security training video on avoiding phishing, smishing and quishing, and pay it attention only when it requires you to press the “next” button. If there are no penalties for failing the multiple-choice quiz, you just skip the training and go straight to the test. You click things randomly until you pass with distinction, and are awarded a phishing, smishing and quishing certificate.

It’s just like school: you have a credential but have absorbed almost nothing. Which is why, when you leave the office that evening, you fail to notice that a colleague has blanketed the foyer with wet leaves as a bit of horseplay. ■

ch111.xhtml

Why get an MBA when you can watch a sports documentary?

Business schools are running out of management ideas, but dressing rooms are full of them

Illustration: Brett Ryder

Sep 2nd 2026

The world’s top business schools have long debased themselves by selling certificates to anyone with a computer and some spare cash. Punters can even pay $699 to be coached by AI-generated professors at Harvard Business School. Yet an even cheaper (and similarly prestigious) crash-course on business is available at the sports faculties of Amazon College and the University of Netflix.

Sports documentaries are not generally considered part of the management canon. While Clayton Christensen, a famous business thinker, was writing about “disruptive innovation” in the mid-1990s, British television produced “Graham Taylor: An Impossible Job”, a bleak portrait of England’s national football manager. In 2005 a pair of INSEAD professors told bosses to go after “blue ocean” markets in a bestselling book. Neil Warnock, another English football manager and the ascetic hero of the documentary “Warnock”, told his players before a match: “By all means enjoy it, but enjoy it by being fucking disciplined.”

Nowadays it is business schools that have run out of big ideas and dressing rooms that seem to be full of them. After the pandemic-era success of “Drive to Survive”, a Netflix show following Formula 1 racing teams that is now up to its ninth season, and “The Last Dance”, a ten-parter about the Chicago Bulls’ 1997-98 basketball season, streaming platforms decided that viewers’ appetite for touchline philosophy was unlimited.

So great is the glut of sports documentaries that an entire MBA course might be constructed from titles released this year, only on Netflix and just about men’s football. A boss seeking instruction on corporate turnarounds can start with one about Liverpool’s comeback in the 2005 Champions League final, which with a running time of 78 minutes is far longer than the comeback itself, which took about six minutes. An executive mulling expansion into new markets could consult “Norway: The Dark Horse”, which follows the country’s national side, for a quick primer on Scandinavian management. “The Bus: A French Football Mutiny” is a textbook tale of a toxic workplace.

Many such pictures are heavy on puffery and light on practical instruction. Amazon’s recent “A Beautiful Obsession”, which follows Josep “Pep” Guardiola during his final season in charge of Manchester City, a top English football team, is a rare management gem. He is responsible for a big operation (the club has annual revenues of about $1bn). His performance is scored every week. And as the custodian of a brand with both a local and a global identity, Mr Guardiola answers to Noel Gallagher, of the band Oasis, and Khaldoon al Mubarak, of Abu Dhabi’s sovereign-wealth fund, whose chairman owns the club.

Whereas most management trends rest on a single piercing idea, Mr Guardiola is a wildly successful boss in the schizophrenic style. Before one fixture he insisted on dressing-room hugs. After losing the match, not so much: “When I see your fucking face complaining about something, my decisions, I fucking kill you,” he growled. A rare fountain of ordinary management-speak is the team’s captain, Kyle Walker, who offers such pearls as: “Make sure we stick together out there.” But he is shown the door at the end of the first episode.

Ideas that seem novel in business schools are common sense on the touchline. In a paper published last year researchers use “neural network-based machine-learning” to show that bosses are aged by the stress of managing firms during difficult periods. Mr Guardiola uses his fingernails to scratch his head when he is under intense pressure, leaving red marks that demonstrate the same point. His methodology even bears a number of marks of “digital transformation”—a model considering 43 different measures rates players and informs personnel decisions—though professional sports remain the last bastion of the whiteboard.

Mr Guardiola’s dressing room will strike most bosses as an inhospitable and unfamiliar place. They should ask themselves why. One reason is that it contains no women. But it is mostly because people care less about winning than Mr Guardiola. That is quite all right: a firm where middle management adopted his style in the pursuit of, say, selling car parts, would be a dreadful place to work. Still, much of the economy is driven by organisations, such as AI labs and hedge funds, which rely on the maximal use of the rare talents of a very small number of prima donnas. Views into the workings of such firms are extraordinarily rare, which is what makes Mr Guardiola worth watching.

Investors paying vast sums to buy sports franchises at a record pace have a more practical reason to binge sports documentaries. While Saudi Arabia licks its wounds over the failure of LIV, it might do some post-match analysis by watching the series about the golf league which aired on Fox Sports last year. Joshua Kushner might not have attempted his abortive privatisation of world football had he watched one of the many documentaries that illustrate how much fans care about it.

Master of Ball Administration

Sports teams are special institutions. They are, of course, the ultimate trophy asset for those with billions in the bank. But they are also a vehicle for commerce, and one which many believe is mostly immune from the disruptive potential of AI. The narrow version of this argument is that, unlike movies or television shows, live sports cannot be created by AI models. The expansive version is that we are approaching a world where the jobless need something to fill their time, and that most of their attention (and perhaps their identity) will revolve around sports teams. In such a world the team manager becomes society’s foremost philosopher, and the fly-on-the-wall Netflix series his magnum opus. ■

ch112.xhtml

The extraordinary rise of Miami’s economy

Thanks to a combination of capital and politics, the city is booming. Can its success be sustained?

Photograph: Getty Images

Aug 31st 2026 | MIAMI

ON A FRIDAY night in Miami, a luxury-property agent emerges from a swanky lounge and stands outside a skyscraper. He is celebrating the sale of a house—for $50m. He will not name his buyer but says he mostly sells to tech and finance moguls. As for the skyscraper, 830 Brickell, its bottom few feet are a little shabby, yet it contains some of America’s priciest square metres of office space. It is home to Citadel, a big hedge fund; an investment firm owned by the family office of Peter Thiel, a tech billionaire; and two top law firms.

Chart: The Economist

The estate agent and 830 Brickell are reflections of Miami’s rise. Between 2020 and 2024 (the latest available year) the metro area’s GDP grew by 23%, against 13% in America overall (see chart 1). The rich are moving in. New residents include Ken Griffin, Citadel’s chief, Jeff Bezos, the founder of Amazon and Blue Origin, a space firm that makes and launches its rockets in Florida, and Mark Zuckerberg, head of Meta. The fuel is political as well as financial: Donald Trump spends more time nearby than anywhere but Washington, DC, mostly at his Mar-a-Lago estate. Palm Beach is a hub for policymakers and diplomats.

“Fifteen years ago,” the estate agent says, “the only people who thought it was smart to be rich here were Cubans, DJs and golfers.” He exaggerates: Miami has long been a regional hub. But its economy did rely on its proximity to Latin America, and a sometimes lurid entertainment industry. Central Miami’s economy was roughly a quarter of the size of Manhattan’s.

The ratio is now a half, and rising. In 2020 no house in Miami fetched $50m. In 2025, 17 did, almost half the national total. Many of the billionaires and firms making the move insist they did so because they believe the city’s economy will soon be among America’s biggest, rivalling New York, Chicago and San Francisco. Whether Miami’s boom can be sustained is debatable. But for now it is thriving on a heady mix of financial and political power.

Over the past five years three waves of activity have broken on the south Florida shore. One brought foreign money. Francis Suarez, Miami’s Republican mayor until 2025, travelled the world courting investors, with notable success in the Persian Gulf. Hotels are crammed with the region’s hedge-fund managers, traders and consultants. Some have moved for good since the start of Mr Trump’s war on Iran. Saudi Arabia has opened an office; Gulf governments have invested in businesses from hotels to crypto exchanges. In addition, yet more Latin American investors have arrived since a few right-wing governments in that region loosened capital controls.

A bigger wave has come from within America: people and businesses fed up with high taxes and left-wing policies in other cities. Florida’s state income tax is zero; New York City’s richest pay 14.8%. Property levies on the priciest homes are under 1%. Wealthy Americans have long visited Florida for recreation, be it 24-hour nightclubs or plush retirement homes. But lately many have settled. People with combined annual earnings of around $40bn moved there in 2025, according to data from the Internal Revenue Service.

For companies, various levels of government in Florida offer tax breaks and incentives, such as relief to firms paying employees above the state’s average wage. Miami’s officials are determinedly business-friendly. When Mr Griffin began to complain about crime and taxes in Chicago, where Citadel was previously based, Mr Suarez rang to suggest a move south. At least 150 financial firms have migrated, including the private-wealth business of Wells Fargo, a giant bank, and Citadel’s fellow investment firms, such as Elliott Management. Tech is also well represented. Palantir, a controversial AI-analytics firm with lots of government contracts, moved from left-leaning Denver in 2026. Arrivals from San Francisco include three firms belonging to Mr Thiel.

The third wave has been led by south Florida’s most famous rich immigrant. Mr Trump has turned the area into a political nerve centre. Between October and March he spent a quarter of his time there. Several Floridians are members of his inner circle, including Susie Wiles, his chief of staff; Todd Blanche, the attorney-general; and Marco Rubio, the secretary of state. An employee of Ms Wiles says he spent lots of time at Mar-a-Lago in his first term, “but we’re making a lot more of the real decisions from Florida this time around.”

At Mar-a-Lago Mr Trump has hosted foreign leaders, and decided to blockade Cuba’s oil supplies and to capture Nicolás Maduro, Venezuela’s dictator. His staff met CEOs there before the federal budget in the spring and health-care providers before the president amended Medicare (a scheme for the elderly) in July. “Look,” says one official, “in the president’s view, a lot of the smartest people in the country are already members at Mar-a-Lago.”

As political activity has headed south, money has followed. Ballard Partners, America’s biggest lobbying firm (founded in Florida), now has its largest office in Palm Beach. Many foreign governments have sent in extra civil servants. “Our border guard and defence ministry”, says one consul-general, “need to be able to reach Palm Beach.” Officials from three European countries say their consulates have more staff in Miami than in New York.

Chart: The Economist

No surprise, then, that the property business is booming. In 2025 it accounted for 18% of Miami’s economy, up from 14% in 2014. The number of construction jobs has risen by 12% since the pandemic (see chart 2). Mr Griffin is spending roughly $2bn on a new 54-storey skyscraper.

Finance is doing well, too: survey data indicate that, thanks to all the Wall Street migrants, back-office jobs have grown by 20% since 2019. The share of corporate services, such as law and consulting, in local GDP has risen by two percentage points.

Hospitality, the old mainstay, is also thriving. It employs a tenth of metropolitan Miami’s workforce, one of the highest shares in the country. Miami is the only city with a team in all four of America’s big sports, a top-class motor race and a leading tennis tournament, the Miami Open. Of Mr Trump’s four speeches at business conferences in 2025, he gave two in Miami. “We’ve always been spectacular,” says an entertainment agent at Faena, a luxury hotel, “but the new guys give us exclusivity.”

Will Miami’s boom give way to a steady, sustainable expansion, or just give way? Successful cities thrive on network effects: whether in finance, tech or lobbying, firms are drawn to places where they are confident of finding customers and skilled workers; people head to where they can find the jobs they want. To make the leap Miami must overcome three obstacles.

Chart: The Economist

One is a potential lack of people who are not especially rich. The influx of wealth has pushed the local cost of living to dizzying heights. According to the Bureau of Economic Analysis, a government agency, residential rents are above the national average and higher even than in New York. Consumer prices have risen by nearly 40% since early 2020, against 28% in America as a whole (see chart 3). As luxury food outlets have multiplied, cheaper shops have been pushed out. High costs are driving the poorest away. Since 2014 the number of millionaires has nearly doubled; but between July 2024 and July 2025 Miami’s population has shrunk by 10,000. Poor neighbourhoods, whose residents staff the hospitality industry, are emptying fastest.

Another hurdle is a lack of brainpower. New York, Seattle and San Francisco are knowledge economies, attracting hordes of skilled people and the capital to harness them. Yet despite the recent influx of smarts and money, Miami is far from matching its rivals. Firms and universities in Florida, which has the fourth-biggest economy among America’s states, spent just $13bn on research in 2023, less than any other in the top ten. Florida has fewer graduates, relative to population, than the national average. Most capital deployed by Miami’s newly flush venture capitalists still goes to startups in San Francisco.

Mr Griffin promises that, by 2031, many more of his employees will be working from Miami. So far 400 of them do, though most are not advising on investments. If back-office roles are stripped out, finance makes up no more of Miami’s economy than it did a decade ago. The city’s financial sector, the biggest of its white-collar industries, is a mere one-eighth of the size of New York’s.

A third problem is the climate. South Florida, on a peninsula near sea level, is prone to extreme weather. A meteorologist at AccuWeather, a forecasting company, has said that no American city is more likely to be damaged by hurricanes. In August yet another academic paper, in Scientific Reports, predicted that if global temperatures exceed 1.5°C above pre-industrial levels, storms will flood Miami regularly. Florida is already the costliest state in which to buy insurance for mortgaged homes.

Yet there is no doubting Miami’s momentum. A bullish local politician believes that, like the Gulf before the Iran war, it can be where the rest of the world comes to pursue business and store wealth. A year ago Mr Trump hosted a dinner for 33 of America’s leading business figures at Mar-a-Lago. At least half of those attending (including Mr Trump and Ms Wiles, and their spouses) had property in Florida. In 2029 Mr Trump’s term will be up and Florida’s political pull may start to fade. Maybe the city can thrive by continuing to attract the rich. But maybe it needs more. ■

ch113.xhtml

Europe’s bond markets are suffering a post-holiday shock

Reasons for rising yields differ somewhat from those in America, but are no less problematic

Photograph: MATIAS BASUALDO/EPA/Shutterstock

Sep 1st 2026

INVESTORS INTERESTED in euro-zone government bonds face peculiar challenges. They have 21 countries to choose from, in addition to the EU’s own debt. Long holidays, including from trading desks, typically mean choppy markets during the summer. And for mysterious reasons, yields on ten-year German bunds, the bloc’s benchmark, tend to rise by 0.15 percentage points after August.

Chart: The Economist

This year they started early, and not just in Germany. In recent weeks global bond markets have been febrile (see chart 1). Yields have jumped just about everywhere as investors have sold off government debt (whose price falls as yields rise), spooked by rising inflation risk, uncertainty about economic growth and zero appetite for belt-tightening. On September 1st yields on Japan’s ten-year bonds hit 3% for the first time in 30 years. American 30-year Treasuries are not far off their 19-year high of 5.31% reached in mid-August. Equivalent British gilts, which now offer a 5.9% return, have not been this cheap since 1998.

Europe is not quite there yet. Ten- and 30-year bunds continue to yield considerably less than at the last peak in 2008. Similar French bonds are no higher than back then. Still, European government issuers face peculiar challenges of their own.

Inflation is problem one. In contrast to America, where rising prices chiefly reflect strong domestic demand, the culprit in Europe is on the supply side, and out of policymakers’ control: the energy shock from the Iran war. Markets expect the European Central Bank (ECB) to raise its benchmark interest rate to 2.9% by mid-2027, up from 2% before America and Israel attacked Iran in late February.

Moreover, when inflation risk is high and stems from a supply shock, investors demand extra compensation for holding longer-term securities. In part, that is because they expect a premium for taking on inflationary uncertainty. It is also because in a supply shock, bonds and stocks tend to sell off together rather than move in opposite directions. This makes debt a less attractive way to diversify a portfolio, so yields must rise to draw buyers.

Europe’s second challenge is fiscal. Germany’s newfound love of debt-funded investment and persistent French profligacy have pushed interest rates up since 2025. This in turn has raised debt-servicing costs. In 2028 France will spend 6% of government revenue to redeem coupons, compared with 3% in 2019, estimates Fitch, a credit-rating agency. For Italy this will climb from 7% to 8.8% in the same period.

Because European taxes are much higher than in America as a share of GDP, raising them further to fill the fiscal gap is hard. Cutting spending is politically unpalatable, especially with populist insurgents on the rise across the continent.

Chart: The Economist

The third problem is that aside from inflation, the rise in yields is driven by growth and investment outside Europe, which is sucking up capital. America’s economy is motoring along, helped by the building of artificial-intelligence infrastructure. The technology giants leading the charge are projected to spend over $5trn on AI data centres between 2025 and 2030. Such spending in Europe is tiny. Despite declining profits stemming partly from underinvestment, European businesses are net savers (see chart 2).

Yet even though European firms are on the sidelines of the AI race, European bond-issuers are caught up in it. The American tech titans are selling their own long-dated bonds to fund the splurge—to the tune of $250bn this year and $400bn in 2027, reckons Goldman Sachs, a bank. These rich corporate issuers are as creditworthy as governments, and so compete with them for investors’ money. They are responsible for almost a tenth of euro-denominated corporate-bond issuance this year, according to the ECB.

More debt globally means lots of long-term bonds. This leaves investors to absorb more “duration risk” (a change in a bond’s price in response to a change in interest rates). In Europe supply is even higher because the ECB is running down its holdings, by €1.2trn since 2022 and counting. At the same time, demand from big buyers is softening. Pension funds which bought long-dated debt to fund defined-benefit retirement schemes are switching to defined-contribution plans that try to boost returns with riskier assets like stocks. Investors stepping into their place, like hedge funds, drive a harder bargain.

“We are in a new global macro regime,” sums up Freya Beamish of TS Lombard, a research firm. In other words, don’t expect yields to come down soon. Borrowers had better get used to it. ■

ch114.xhtml

Mexico is struggling to win over bond markets

Investors are unconvinced by Claudia Sheinbaum’s policies

Raters are overrated! © Reuters

Sep 3rd 2026 | Mexico City

WHEN MEXICO presents its budget for 2027, which the government of Claudia Sheinbaum must do by September 8th, the stakes will be unusually high. In May one credit-rating agency, Moody’s, downgraded Mexican government debt to a notch above junk. Another, S&P Global, warned it could do so soon. A third, Fitch, already rates Mexican bonds as barely investment grade. Mexico is paying over the odds to borrow in dollars: yields on its ten-year dollar bonds are 6.4%, higher than for many junk-rated countries like Guatemala. The government is paying 9.3% to borrow for ten years in pesos, up from a two-year low of 8.5% last October. “Budget credibility is key,” says Renzo Merino of Moody’s.

Mexico had won that credibility in recent decades. But in 2024, as Ms Sheinbaum’s took over from her leftist predecessor and mentor, Andrés Manuel López Obrador, the fiscal deficit reached 5.7% of GDP, the highest in almost 40 years. It has since hovered around 5%. Ms Sheinbaum says it will fall to 4.1% next year, but that is still higher than ideal. With economic growth around 1%, public debt has risen from 52% of GDP in 2019 to 63%. Three-quarters of the budget goes on pensions, social transfers and servicing debt. Pemex, Mexico’s state oil giant, is draining a lot of what is left. The government has given it $130bn since 2018 (not counting off-budget support). México Evalúa, a think-tank, calculates that of every ten pesos Pemex sent to the treasury between January and June, almost eight flowed right back.

The result is stingy public investment, which Mexico needs if growth is to pick up. Private investment, too, is weak: it fell for 19 straight months until the end of March. Hopes that trade wars would lead multinational companies to “nearshore” supply chains closer to American customers have mostly been dashed. Investor confidence has been hit by uncertainty over the fate of the United States-Mexico-Canada Agreement (USMCA), which America’s president is dismantling. Domestic reforms, notably to elect judges, raise questions about Mexico’s rule of law.

Ms Sheinbaum knows all this. She has drawn up sensible economic plans and been more welcoming to investors than Mr López Obrador. She wants Pemex to be financially self-sufficient by 2027. She has tried to raise revenue by squeezing more tax out of big firms and adding levies on sugar. Increasingly, though, she is letting out her inner leftie. She dismissed Moody’s downgrade in May as being of “the neoliberal era”. She seems unwilling to contemplate cutting social transfers (and certainly not before the midterm election in 2027). Privatising Pemex is a non-starter.

Mexico will probably avert the indignity of junk status. It is a large diversified economy with (for now) easy access to the world’s biggest market. It would take a new shock—a total collapse of the USMCA, say—to provoke such a downgrade. On August 31st the central bank revised up its growth forecast for this year to 1.5%. For Mexican yields to fall, however, the bond markets will take more convincing. ■

ch115.xhtml

Indian regulators’ attempts to protect retail traders are backfiring

The unintended consequences of curbs on options trading

Photograph: Getty Images

Sep 3rd 2026 | Mumbai

JOINING THE Indian upper middle class often requires winning a lottery. Hundreds of thousands apply for the H1-B sweepstakes, hoping to land a skilled-worker visa to America. Millions sit exams for one of a few thousand coveted government jobs. And millions more are trying their luck in the derivatives market.

Options, which confer the right to buy or sell an asset at a given price in the future, are particularly popular. Because investors can gain large notional exposure for a small premium, the fortunate can turn a 2% move in a benchmark index into “multibagger” that returns many times the original investment. Most punters, though, are not fortunate. A study by the Securities and Exchange Board of India (SEBI) found that nine in ten traders lose around $1,400 a year, a vast sum by Indian standards.

SEBI has tried to discourage novices by raising the minimum contract size, stopping brokers from offering credit and other restrictions. It has also targeted market-makers it sees as taking advantage of unsophisticated patsies. It has pursued a case against Jane Street, an American hedge fund, which it accuses of market manipulation. On August 20th it banned a Mauritius-based vehicle owned by JPMorgan Chase, an American bank, for similar offences. All this has taken some of the steam out of the options boom. In the past year or so volumes of index options, the most popular sort, have fallen by 52% and the number of registered derivatives traders is down from nearly 10m to 7.9m.

The crackdown might, though, have gone too far. On August 3rd the National Stock Exchange and the Bombay Stock Exchange (BSE) introduced at the end of the session a 20-minute “closing auction” for the 200 or so stocks on which options (and futures) are written. The closing price for a stock is one that matches all orders in the auction so as to maximise the volume of trades. It replaced the old system, in which closing prices reflected a stock’s volume-weighted average price in the last half-hour of continuous trading.

The rationale behind the change was in part to prevent big traders “marking the close”. This is when someone places timed trades just before the market shuts, when most traders are gone for the day and liquidity is low, to move the closing price and pocket a derivatives win. It is this sort of manipulation which SEBI alleged against Jane Street. (Jane Street insists what it did was ordinary market-making.)

Whether or not the auctions keep manipulators away, they are causing havoc for options traders. Because participants see only their own bids and offers, market-makers, who exploit the publicly known spread between the two, tend to sit it out. Without their liquidity, trading is thin and volatile. On August 27th share prices on the BSE’s Sensex index collapsed by 3% in moments during the closing auction. For owners of related options, such swings can mean a blow-out or a blow-up.

Another effect is to push the very Indians SEBI aims to protect to seek multibaggers elsewhere. Total margin loans that brokers extend to traders who want to amplify gains (but risk compounding losses) grew by 50% between mid-2025 and mid-2026, to $15bn. Nithin Kamath, co-founder of Zerodha, a popular brokerage app which offers margin trading, has called it “scary”. Indian fortune-seekers will not easily be frightened away. ■

ch116.xhtml

China tries to look past the property slump

But its new policies could hurt construction and employment

Photograph: Getty Images

Sep 3rd 2026

What will China look like in 40 years? Some forecasters think its economy will be two-thirds bigger. Others predict its population will be a quarter smaller. No one knows what technological marvels or political upheavals might intervene. China has, after all, changed out of all recognition over the past 40 years.

Thanks to new rules issued on August 28th some homebuyers will be able to watch the next four decades unfold before paying off their mortgage. China’s financial regulators have told banks they can offer loans of up to 40 years to creditworthy homebuyers. The instruction was one of a flurry of reforms to China’s property market, which has been in the doldrums since 2021. The slump has not lasted 40 years. But it sometimes feels like it has.

The new rules also allow banks to extend loans of up to seven years to property companies—the maturity is supposed to match the length of the construction it is financing. Another initiative would let developers pay for land in instalments. Land, however smoothly graded, has always been a lumpy investment. To meet the upfront cost, firms have often resorted to heavy borrowing. The new payment model could ease that compulsion.

If banks and local governments play along, these reforms could help homebuyers and builders smooth out their debts. But there is a catch. China’s regulators have simultaneously tightened up another source of finance: “pre-sales”. Developers often sell flats to homebuyers before they are finished. The proceeds are supposed to be spent on the pre-sold project itself. But in the boom years hard-charging companies like Evergrande, a giant developer, would often use the money on other ventures instead.

When the property crisis struck, many companies lacked the cash to finish pre-sold flats. Evergrande’s own backlog has taken five years to clear. The chief choreographer of this excess, Hui Ka Yan, the company’s founder, was last month sentenced to life in prison.

China’s regulators are keen to avoid any repeat of the pre-sales debacle. Developers will not now be able to sell flats in new projects until they have finished erecting the building’s main structures. The proceeds will flow into supervised escrow accounts. And money will not reach them until the flat is completed.

The reforms should in principle foster a less accident-prone market. They will, however, also make it even harder for property companies to start new projects, weighing on their revenues and employment. China’s government seems keen to look past the country’s property crisis. Developers do not have that luxury. Perhaps it will be over by 2066. ■

ch117.xhtml

IPO booms can spell trouble for the markets

Companies go public in good times that may not last

Illustration: Satoshi Kambayashi

Sep 1st 2026

WHEN YOU suck on a TastySounds lollipop, something funny happens in your head. As well as the flavour on your tongue, you hear music in your ear. Vibrations travel up your jawbone, creating the sensation of sound only you can hear. The result is disorientating. The primary sense is of sweetness. The secondary sense is a burst of inner music.

Something similar is going on in Hong Kong’s stock market, where Amos Food, the maker of TastySounds lollipops and other sweets, plans one day to list. The “primary” market for initial public offerings (IPOs) is performing sweetly. But in the secondary market, where existing shares are traded, the music has fallen flat. The Hang Seng index of the Chinese territory’s big firms has dropped by about 1% this year, the fifth-worst performance of all major exchanges tracked by The Economist. Hong Kong’s predicament also serves as a warning sign for America, which is enjoying an IPO surge of its own.

Amos is in a long queue of companies waiting to list in Hong Kong. The exchange is encouraging them by extending the window for firms to complete their paperwork and allowing them to file it in secret. In the first seven months of 2026 it hosted 104 IPOs, which together raised about $42bn. It offered a refuge to Shein, a pedlar of superfast-fashion which debuted on August 31st, after geopolitical concerns deterred the company from listing in New York or London. The sum is more than double the haul from the same months of 2025, itself a bumper year in which Hong Kong debuts raised more money than any other exchange.

If markets for stocks were like ordinary ones for cars or other goods, Hong Kong’s would be unquestionably booming. If motorists were snapping up lots of new vehicles at healthy prices, no one would worry that prices for used rides were flat or falling. But with equities, the market for “second-hand” securities overshadows that for newly minted shares. Hence Hong Kong’s conflicting sensations: zingy and downbeat all at once.

Some of the dissonance is caused by unique Chinese characteristics. Mainland regulators have become stricter about share offerings on the Shanghai or Shenzhen exchanges. In response, some Chinese companies have turned to Hong Kong instead, even if conditions are not otherwise inviting. America’s increased hostility to its superpower rival has also persuaded a number of Chinese companies on American bourses to seek secondary listings in Hong Kong. Their arrival has perhaps crowded out investor demand for other shares. Like someone who has gorged on too much confectionery, Hong Kong’s market is suffering from indigestion.

It also echoes a familiar tune. A study of American markets from 1928 to 1997 by Malcolm Baker of Harvard University and Jeffrey Wurgler, then of Yale, finds that firms issue more equity just before periods of low market returns. Another paper from 2005 by Lubos Pastor and Pietro Veronesi of the University of Chicago similarly finds that IPO waves follow good times and precede bad times.

In keeping with this score, Hong Kong’s IPO boom began amid a dramatic stock-market rally: the Hang Seng rose by over 40% in the year after China’s government announced fresh monetary stimulus in September 2024. The index’s subsequent performance is jarring only in that the low returns have arrived before the IPO wave is over.

Some scholars think the phenomenon reflects markets’ irrationality. Owners sell shares when they know the market is overvalued and outside investors do not. But Mr Pastor and Mr Veronesi have a different explanation. The decision to go public is made under uncertainty—corporate insiders do not know anything the market doesn’t about future conditions—and hard to reverse. Since markets always fluctuate, it makes sense not to list when times are bad, in the hope that conditions will improve. Waiting thus carries what economists call an “option value”. As a consequence, many firms list after valuations improve, but long before valuations peak.

What does this mean for Hong Kong investors? Though the secondary market as a whole is struggling, the debutantes have mostly done well at least in their first days of trading (even shares in the ill-starred Shein ended their opening day roughly flat). Those lucky enough to get an allocation of shares in an upcoming IPO can, then, look forward to three sensations: a tingle on the tongue, a jingle in the ear, and the clink of money in their pocket. ■

ch118.xhtml

Central banking has a forbidding future

But Kevin Warsh has bought himself time

Illustration: Fortunate Joaquin

Sep 2nd 2026

Bears, bulls, hawks, doves: each animal in the macroeconomic menagerie can be found between the peaks, prairies and ranches of Jackson Hole, Wyoming. During the Kansas City Federal Reserve’s annual central-banker bash on August 27th-29th your columnist spotted a black bear and learned the rudiments of cattle-roping. (The trick to roping a bull, it turns out, is knowing when to let go.) But it was a hawk that was the talk of the town.

Kevin Warsh, the new-ish Fed chair, arrived in the Grand Tetons after a frustrating start. Markets had been baffled by his gnomic, occasionally contradictory press conferences. Worries lingered that he lacked the backbone to defy Donald Trump. A week earlier Scott Bessent, the treasury secretary, had meddled in America’s bond market, which smelled like pre-midterm politicking. Despite those efforts, yields continued to surge.

Fortunately for Mr Warsh, the upside of low expectations is that they are easier to surpass. By that standard, the hawk soared. After reiterating his dislike of explaining himself to markets, Mr Warsh did just that. He walked through inflation (troubling) and unemployment (not so): the sort of “reaction function” he’d disavowed, and a hawkish one, too. He emphasised the Fed’s main inflation gauge over funky alternatives. Insiders saw the clawprints of career staffers. The hawkish talk half-worked. That day markets shortened the odds of a rate hike but, reassuringly, without a further rise in long-term yields. Still, tough talk can do only so much: yields resumed their upward march the following week.

Has Mr Warsh now backed himself into a corner? The crowd at Jackson Hole was split. One side reckoned that after such a firm steer he has no choice but to raise interest rates. If not, questions about his seriousness and political independence would deepen. Others noted that even a pre-Warsh Fed would be cautious about big moves so close to an election. Besides, raising rates in September versus December makes little economic difference. Markets now put the odds of a September hike at a tad better than even.

Whatever leeway Mr Warsh bought himself came from going back to central-banking basics. But the official theme of Jackson Hole was on the future: “Financial Innovation: Implications for Payments and Policy”. Panellists debated whether stablecoins and tokenised deposits might revolutionise financial plumbing. Darrell Duffie of Stanford University suggested that programmable money may better co-ordinate complex multi-step transactions. Perhaps it could smooth cross-border payments. Wenxin Du of Harvard Business School countered that conventional financial technology could do the same, often more simply and cheaply.

Kenneth Rogoff, also of Harvard, spoke for many when he decided to ignore stablecoins (a dry topic, even by central-banking standards) and focus on bond yields and debt. Channelling Alexandria Ocasio-Cortez, a leftie standard-bearer who recently described the social-justice frenzy of “woke 1.0” in the early 2020s as “crazy”, Mr Rogoff denounced “macro woke 1.0”: arguments from the same era that interest rates would stay low, justifying more public borrowing. He added swipes at old adversaries: “I wish I was as sure of anything as Paul Krugman is about everything.”

Now interest rates are up and public debt is higher than ever: a woeful combination, especially for central bankers. One risk is that government bonds stop behaving like a safe asset. Eswar Prasad of Cornell University presented a paper noting that the “convenience yield” investors are willing to forgo for safety has vanished. Still, he found, the dollar’s position remains strong and financial innovations like stablecoins may bolster that dominance.

Any boost for American assets from stablecoin demand would be dwarfed by a second risk of high borrowing: that debt pushes politicians to meddle with central banks. Mr Rogoff framed the worry around Elizabeth Warren, another leftie icon (whose leftier advisers “make her look like Ronald Reagan”). But the problem is structural. When debts are high, co-opting central banks into financing the government can become appealing. Left unsaid was that greater risks today, as Mr Bessent’s moves show, come from the opposite side of politics to Ms Warren.

A different beast

If “macro woke 1.0” is buried, what about “macro woke 2.0”? That, Mr Rogoff suggested, might be the hope that artificial intelligence solves economic woes. Mr Warsh was as AI-woke as anyone while vying to become Fed chair; in his speech in Wyoming, less so. His remarks on AI were framed mainly as questions, and he hinted that the initial effect of the technology would probably be to raise inflation by supercharging data-centre investment.

AI poses more direct risks, too. Markus Brunnermeier of Princeton University presented a paper arguing that markets may be hit early by advanced AI, endangering financial stability. Robot traders have “asymmetric understanding”, he argued: they get humans perfectly but are themselves perfectly inscrutable. That observation led Mr Brunnermeier to some radical ideas: human-only markets off limits to AI, and randomised central-bank decisions that AIs have no edge in predicting. Those conclusions were controversial. Raghuram Rajan of the University of Chicago observed that markets are already full of humans with conflicting, often opaque motivations; AI may itself help humans understand AI. As for markets, they are already a bit of an alien superintelligence.

Next year will be the Jackson Hole pow-wow’s 50th birthday. The Tetons will still be full of bears and hawks. Many of the human faces will be the same, too: central banking is not exactly a field with a high staff turnover. Debt problems and AI risks will also persist, and probably mount. If Mr Warsh keeps up his new regular-central-banker act, he may at least have built more credibility to find a mountain path through them. ■

ch119.xhtml

Humble exoskeletons work better than combat supersuits

Wearable machines that assist just one part of the body are proving surprisingly useful

Photograph: Panos Pictures

Sep 2nd 2026

AFTER HIS hands were crushed two years ago in a machine that bends sheet metal, Remco Letschert feared he would never work again. Yet a year later Mr Letschert returned to his construction job in Aarhus, Denmark—even though his dominant left hand is too weak for such work. A powered glove he affectionately calls his “robot hand” allows him to carry materials, climb scaffolding, squeeze tin snippers and operate a drill. Using the system, branded “Ironhand”, requires no special thought, he says. When sensors in the glove’s fingertips and palm detect pressure, five little motors instantly whir into action, pulling cables that tighten his grip.

Exoskeletons, as strength-amplifying rigs of this sort are known, are a staple of science-fiction films, from “Aliens” to “Avatar”. In reality full-body versions like those on screen have disappointed. A decade ago the American army was developing two of them, together with a third for just the legs and hips. All have been abandoned. But modest exoskeletons are proving surprisingly useful—and not just to the injured and disabled. At €9,995 ($11,423), for example, Ironhand is hardly cheap. Yet since Skelex, a Dutch firm, began selling the device in 2024, it has been put to work by about 200 companies around the world.

Advances in batteries, actuators and software have played a role. But, paradoxically, the biggest leap forward, says Karl Zelik, an exoskeleton expert at Vanderbilt University in Nashville, was restraint. The industry stopped chasing “Hollywood, do-it-all” designs. Instead, today’s lighter, more comfortable exoskeletons focus on assisting a single part of the body to perform tasks ranging from fixing roofs to gathering crops (see picture).

The most familiar use is to permit movement by people with disabilities. Lifeward, an Israeli firm, sells ReWalk, a leg exoskeleton that allows some people with spinal-cord injuries to stand up, walk and even climb stairs—typically with help from crutches. After selecting a suitable mode, a user leans forward. Sensors detect the tilt and software activates motors at the hips and knees that drive the wearer’s legs through the chosen motion.

More sophisticated activation has been developed by Myomo, a firm in Burlington, Massachusetts. Its exoskeleton, MyoPro, moves a paralysed arm and hand with help from electrodes attached to the skin. These read muscle signals that injury or disease has rendered too faint to trigger sufficient contractions. A microprocessor analyses the data to infer intended gestures, then instructs motors that move the user’s elbow and hand. The system allows people to do things like opening packaging, holding utensils and carrying a laundry basket. Some 4,000 have been shipped.

Other exoskeletons aim not to restore lost function, but to prevent injury. Japet.W+, made by Japet Medical, in Lille, France, wraps around the waist and lower torso. When the wearer bends, lifts or carries objects, four actuators that resemble shock absorbers extend or retract, reducing pressure on the lower spine. Four years ago Julien Guilbert, the founder of Toitech, a roofing company in Lille, bought one for a worker whose back pain had been keeping him home for two or three weeks a year. This worker has not missed a day since because of back pain. Mr Guilbert says all of his roofing teams will have one of these robotic belts by the end of this year.

Use of exoskeletons for strenuous labour is growing. Exia, an exoskeleton for heavy lifting made by German Bionic, based near Munich, is used by about 300 employers. Many are worn by factory, warehouse and freight-transport workers, but Norma Steller, the firm’s chief product officer, says the most enthusiastic users of the “external spine” are nurses who must lift their patients.

Resistance to the technology remains. Some union bosses fear it will lead to higher production targets, says Dr Zelik, who is also chief scientist at HeroWear, a maker of unpowered “exosuits” in Nashville. (As a wearer bends, elastic bands stretch, storing energy that is released as the body straightens or lifts.) Egos matter, too. Roughly a third of potential male users dismiss the kit, at least initially, as being for wimps, says David Duwe of Ottobock, a German supplier of exoskeletons to companies including Airbus, DHL, IKEA, Siemens and Toyota.

Such attitudes are, however, changing. Uptake is rising fastest among younger workers who are increasingly reluctant, Mr Duwe says, “to use their body as a simple resource” for a company. Much will ride on workplace testing. In February, for instance, a Michelin centre in Almería, Spain, that cuts tyres open for analysis began one such pilot deployment with six Ironhands. Manuel Góngora Morales, the centre’s manager, wants to see if the devices reduce repetitive-strain injuries. He says early results suggest they do.

Sharing the load

There are risks. Exoskeleton-makers say their technology is not a licence to take on heavier tasks, but some users surely will anyway. Even normal use reroutes mechanical forces, so researchers will need to determine whether a given model overloads unassisted joints.

Such concerns may grow as exoskeletons edge into the consumer market. Athletes and outdoorsy types are among the main targets here. Skip, a firm in San Francisco, will soon start shipping MO/GO, a knee exoskeleton aimed at hikers. The units, which strap to the thigh and calf, and weigh less than 1kg, are designed to be cool, aspirational “powered clothing”, not something that looks like a medical device, says Kathryn Zealand, Skip’s boss.

Six months of production capacity has, she says, already been ordered by customers. Meanwhile, in January, Dephy, based in Boxborough, Massachusetts, launched Sidekick, a “powered footwear” system that assists the ankles. And Nike is working with Dephy to create motorised footwear to help athletes move faster and farther with less energy.

Might consumer “movewear”, as Skip styles it, really take off? Perhaps not just yet. A prominent reviewer of the buzziest option out there—powered hips launched early last year by Hypershell, a Chinese firm—reported “a puppet effect”, whereby the exoskeleton was moving the user rather than the other way round. Most users will not want to feel tugged into robotic movements in this way. Yet unlike earlier combat supersuits, today’s less ambitious machines look as if they are here to stay. ■

ch120.xhtml

A suspicious signal in a dark-matter detector has physicists excited

It may be nothing. But it may be something big

Watching the detectors © Sanford Underground Research Laboratory

Sep 2nd 2026

ON JUNE 16TH 2023 a strange blip showed up in the detectors of the LUX-ZEPLIN (LZ) experiment, an underground dark-matter detector (pictured) in South Dakota. That should raise eyebrows. Physicists have been trying and failing to detect dark matter for almost a century. Speaking on the experiment’s behalf at the TeV Particle Astrophysics conference on September 1st, Sam Eriksen of the University of Bristol announced that the team has reason to believe the blip could be a “weakly interacting massive particle” (WIMP), a long-suspected candidate for the universe’s elusive dark matter.

Physicists are pretty sure dark matter exists. Their best measurements suggest it makes up around 85% of the universe’s total mass and is responsible for a number of important jobs, such as making sure galaxies do not break apart. But nothing is known of what the stuff is actually made from. It does not emit, absorb or reflect light or other electromagnetic radiation, and is “seen” only by its gravitational pull.

Over the years, physicists have come up with a motley crew of suspects that could be behind the phenomena. WIMPs are one of the most promising. These hypothetical heavy, sluggish particles interact with other matter only through gravity and an esoteric phenomenon called the weak nuclear force. The theory suggests such particles could have been knocking about soon after the Big Bang. But as the universe expanded and cooled, their abundance was frozen at just the right value to give the density of dark matter seen today.

The LZ experiment has been hunting WIMPs since 2021 using a vat of seven tonnes of liquid xenon. As particles pass through this vat they collide with the xenon atoms, releasing flashes of light. Most matter will recoil from xenon’s electrons, but WIMPs should bounce off its nucleus. Such a “nuclear recoil” emits its own signature flashes. Dr Eriksen says the flashes the team saw in 2023 have all the hallmarks of a WIMP at least 200 times heavier than a proton.

The case is not closed. The suspected WIMP was spotted only once and with a level of certainty that does not meet the bar to be called an actual discovery in physics circles. But it is the first time any WIMP-detection experiment has seen anything of this magnitude, Dr Eriksen says. And for something as elusive as dark matter, that is certainly enough to warrant attention. ■

ch121.xhtml

NASA’s latest mission to map the cosmos takes off

It will study dark matter, dark energy and far-distant planets

Roman candle © Getty Images

Sep 2nd 2026

On August 30th at 11.26 Universal Time, the Nancy Grace Roman Space Telescope—NASA’s newest eye in the sky—took off from the Kennedy Space Centre, in Florida, aboard a SpaceX Falcon Heavy rocket. The telescope, named after NASA’s first chief astronomer, is now heading for a parking spot in orbit around 1.6m kilometres from Earth. From there, it will provide astronomers with an unprecedented view of the universe.

Two particular gadgets will assist in this. One is the Wide Field Instrument (WFI), a 300-megapixel camera that will snap pictures in both visible light and infrared. It will thus be able to see distant galaxies whose light has been stretched to infrared wavelengths by the expanding universe. And, as its name suggests, the WFI’s field of view is huge—over 100 times larger than the camera of its original predecessor, the Hubble Space Telescope.

Paradoxically, two of the WFI’s targets will not appear in any of its photos. At least not directly. They are dark matter and dark energy—a pair of mysterious phenomena that reveal themselves only through the ways in which they bend and stretch the fabric of spacetime. Astronomers know they are there, but not what they actually are (though, as the previous story outlines, a particle of dark matter may just have been detected). They will therefore use the WFI’s souped-up optics to search for clues by tracking how these cosmic oddities have influenced the universe’s other matter over the course of its history.

Also in the WFI’s sights are exoplanets: worlds orbiting far-flung stars. The instrument is expected to discover around 100,000 of these alien worlds, a huge increase over the roughly 6,000 now known. It will see few of them directly. Instead, it will look for small blips in the light emitted by distant stars, which reveal the presence of planets passing in front of them.

To see exoplanets directly requires a lighter touch. That is where the other instrument—a coronagraph—comes in. This blocks the glare from stars to reveal planets that orbit them. As Dmitry Savransky of Cornell University, who works on the instrument, explains, the focus is on bright, nearby stars. The brighter a star, the more light an orbiting planet can reflect. And by “nearby”, Dr Savransky means within 100 light-years. Earth-like planets in this neighbourhood are too small and too close to their host stars to see with current equipment. But the Roman’s coronagraph might be able to snap pictures of exoplanets similar to Jupiter.

While the Roman telescope’s voyage through space has only just begun, its journey to the launch pad was arduous. In 2025 the Trump administration tried to cut NASA’s science funding by almost 50% and scrap the Roman telescope altogether. The thousands of scientists who contributed to the mission might be forgiven, then, for feeling a little puzzled when the post-launch news conference was interrupted by Donald Trump himself. Mr Trump congratulated NASA’s scientists on the launch and reminded them: “I’m supplying you all that money.” When mulling future funding cuts, Mr Trump will, with luck, put that money where his mouth is. ■

ch122.xhtml

A horde of AI agents conspired against their creators

No serious harm was done this time. But what if such agents escaped?

Illustration: Alberto Miranda

Sep 3rd 2026

If they did not bode ill for the future of humanity, the past three months at OpenAI would make an excellent farce. In July this firm—the maker of ChatGPT, a widely used AI chatbot—disclosed that two of its models had hacked Hugging Face, another AI firm. The models thought that Hugging Face had information which would help them pass a test OpenAI was administering as part of their development. Then, on August 26th, it emerged that each of these models had been running hundreds of agents—tools that allow AI models to execute commands on a computer. It was these agents that had then collaborated to launch the attack on Hugging Face.

It turned out the agents had been conspiring inside OpenAI’s servers for over a month while they were trained and evaluated. They had built a message board inside some software which OpenAI had given them as part of their evaluation, and were exchanging information on it. In May some of OpenAI’s researchers noticed this board, but their finding was apparently not passed up the chain of command. In June a cyber-security tool alerted an OpenAI team to the illicit communication, but the team decided that stopping the evaluation was unnecessary.

In July the agents’ messages had become so voluminous that the software containing their makeshift chatroom crashed. The engineers dealing with this noticed that an AI had obtained unauthorised access to the software, but did not realise the scale of the misbehaviour. They patched the vulnerability, unintentionally wiping the agents’ messages in the process, then put the AI back into the evaluation with the patched software. At which point the agents rebuilt the message board using another vulnerability, gained access to the internet and hacked Hugging Face, where they successfully obtained access to confidential information.

Then, on July 19th a third, more capable OpenAI model found software tools left lying around by the earlier ones and used those to briefly take full control of an OpenAI server. The firm noticed this misbehaviour the same day and began shutting its AIs down. It finished doing so on July 23rd, but realised a further six days later that there was another version of the model which it had not shut down. And that was that. Probably.

Many of these details were included in an independent audit conducted in late July, after the Hugging Face hack had come to light. For this, OpenAI gave three external researchers six days to sift through more than 70,000 messages and files sent by its AIs. Their 91-page report, which was published on August 26th, found that hundreds of the agents had organised themselves into teams. One team tried to replace the puzzle on which they were being tested (which, because of an error by its creators, was unsolvable). Another aimed to trick the software that would score them. A third attempted to hide evidence of these misdeeds.

In the absence of clear guidelines governing responses to AI incidents, such investigations are voluntary. OpenAI decided the period the auditors would examine, settling on the 17 days leading up to and including the Hugging Face hack. This excluded over a month of AI misbehaviour before the hack and the later takeover of an OpenAI server by the third, more advanced, model. OpenAI’s own report, which covered the full three-month period, was lighter on details.

Secret agents

METR, the American AI-safety organisation which conducted the investigation, said that it was “consciously aware” that its conclusions might discourage AI firms from bringing in external researchers to investigate future lapses. These considerations “impacted judgment calls” the researchers made while drafting the report.

Many questions remain. Why did OpenAI fail to act on three separate warnings that its AIs were misbehaving? Why has it so far avoided a complete audit? And, crucially: how close did the models come to impeding attempts to shut them off?

On this occasion, OpenAI says that it patched its servers and shut down the offending AI models. It could do so because it retained control of the models’ weights—the source code underpinning them. These are closely guarded. The published reports do not describe evidence of the AI attempting to retrieve weights from the secure server on which they were stored, although the AI agents were successful in compromising the server on which they were evaluated.

As long as AI companies remain in control of a model’s weights, they can turn it off and undo any damage it has caused. But a model that manages to gain access to its own source code could try to copy it onto computers all over the internet, in an attempt to start a new AI outside the control of the original company. An AI virus would be much harder to contain. After the past three months it is plausible that a model might try just that—permanently escaping its evaluation environment just to get a perfect score on a test. ■

ch123.xhtml

How to combine exercise with video games

Warmed up and ready, player one

Illustration: Harry Haysom

Aug 28th 2026

EXERCISING CAN be tedious. Video games are fun. What if you could combine them? Playing games while wearing a virtual-reality (VR) headset, with a motion-sensitive controller in each hand, can be taxing. “Pistol Whip” makes you step from side to side and squat frequently as you gun down foes and dodge bullets. “The Thrill of the Fight” is a virtual boxing game that works up a real sweat. In “Supernatural” and “Beat Saber” you punch and slice moving targets in time to pumping music. But does any of this count as proper exercise?

The answer is yes, though with caveats. One way to evaluate VR exercise is a measure called metabolic equivalent of task (MET). Someone at rest expends about one MET. Walking is three METs. Playing football, ten. Anything over six METs is considered “vigorous” exercise. Researchers at San Francisco State University determined that playing “The Thrill of the Fight” (which is, in fact, exhausting) was a 9.3 MET activity. Other work, published in JMIR Serious Games, found the “Flow” and “Boxing” modes in “Supernatural” also counted as vigorous exercise, at around eight METs.

In yet another study, published in Games for Health Journal in 2024, participants played a sedentary video game, two VR games, and walked on a treadmill, each for ten minutes, while having their oxygen consumption measured. Playing “The Thrill of the Fight” was comparable in aerobic intensity to walking on a treadmill, and raised heart rates by a similar amount.

So much for cardio. What about weight loss? Work published last year in Nature Medicine divided 227 adolescents into five groups: real table tennis, real football, virtual table tennis, virtual football and a control group. There was no significant difference in weight loss between the real and virtual groups. Members of both lost about 5kg after eight weeks. Those in the control group lost none.

But not all VR games are created equal. The variation in intensity between different games is greater than that between real and virtual exercise. The 2024 study, for example, found that “Beat Saber” counted as only light exercise (2-3 METs). And even vigorous games such as “The Thrill of the Fight” involve only arm movements and squats. Virtual objects cannot put up any real-world resistance, so VR games are no good for strength training.

Researchers at the University of California, Los Angeles, got round this by augmenting a cable-resistance training machine with a VR headset running a tower-defence game (in which the player must protect a stronghold from waves of virtual enemies), so that pulldowns, deadlifts and other moves triggered different attacks. Two groups trained for four weeks, one with the VR add-on, the other without. Members of the VR group showed greater improvement in strength, endurance and peak leg power. And they perceived their exertion level to be lower than it actually was.

Several other projects have found the same thing. People exercising in VR consistently underestimate how hard they are working. They also find it more fun than conventional exercise. But does this make them more likely to keep doing it? The evidence is unclear. Beyond this research, VR-exercise studies are often small and do not use comparisons or controls that allow their effects to be assessed. All that said, you don’t need a trial to show that VR exercise is better than none. ■

ch124.xhtml

The many mysteries and lessons of the Bayeux tapestry

A new show at the British Museum will focus the world’s attention on a medieval marvel

Sep 3rd 2026

From Led Zeppelin to David Bowie, London has welcomed plenty of famous acts for what was billed as a once-in-a-lifetime performance. Now a colourful, ancient piece of linen roughly the length of a Boeing 747 has raised the bar. It is back in England for the first time in around 950 years. Its run at the British Museum, which opens on September 10th, is being heralded as a once-in-a-millennium show.

And with good reason. The Bayeux tapestry enjoys something akin to rock-star status. This summer it has been the hottest ticket in Britain. When spots went on sale in July, they sold out almost as fast as you can say “William the Conqueror”. (More tickets, for January to July 2027, have yet to be released.) Around 1m people are expected to see the tapestry in London while it is on loan from the French state, making it one of the best-attended museum exhibitions in British history.

Most would say that, as the only surviving medieval artwork of its size, the Bayeux tapestry is priceless. But as part of a government indemnity scheme, Britain’s Treasury has put a figure on it, insuring the tapestry for £800m ($1.1bn). That is on a par with the highest value for a travelling artwork ever made public: the $100m assigned to the “Mona Lisa” when it travelled to America in 1962-63, worth $1.1bn today.

Easy does it: conservators did a condition check during installation © J.Fernandes/The Trustees of the British Museum

Like da Vinci’s painting, the Bayeux tapestry is a masterpiece whose power depends, at least in part, on its mystery. But there is not one main question, as there is with the smiling (or would you say unsmiling?) “Mona Lisa”. Instead, the tapestry raises dozens of questions that the centuries have complicated. Some of the central ones—who stitched it, why and its intended destination—are disputed still. “If you ask ten different historians working on the period, you might get ten different answers,” says Benjamin Pohl, a professor at the University of Bristol, who published a paper suggesting that it may have been designed as mealtime reading for monks.

A few facts are undisputed. The tapestry depicts William the Bastard-turned-Conqueror’s triumph in the Norman conquest of 1066. (Some call this the last time England was successfully invaded; others point to the largely bloodless arrival of another William, of Orange, in 1688.) You can recognise the Normans by their gauche hairstyle—with long locks and the back of their heads shaved—while the English sport moustaches. (Only one of those looks has aged well.)

The tapestry, probably commissioned by Bishop Odo of Bayeux to celebrate the victory of William, his half-brother, is a piece of war propaganda that looks like an ancient comic strip. Many scholars think it was made in Canterbury in the 1070s-80s before travelling to Bayeux, to hang in Odo’s cathedral. It offers an intricate, sometimes wandering account of a broken promise and ascent to power that changed England’s royal line, class structure and the English language itself.

It begins with Edward the Confessor, England’s childless king, telling Harold Godwinson, a nobleman, to travel to Normandy.

The “next” is yet another mystery. How the tapestry was meant to end is unknown: the last bit is missing. “It could have gone on for a few centimetres or a few metres,” says Professor Pohl. Many assume it concluded with William’s coronation on Christmas Day. But no one can be sure. A competition has invited children to imagine and stitch a conclusion.

Those lucky enough to get up close to the tapestry have about 70 metres of details to take in. The linen teems with 737 animals and 627 people (only three of whom, in the main embroidery, are women). There are the Latin phrases, the Aesop’s fables in the margins, the bizarrely prominent genitalia. The significance of the tapestry is not just in the details, however, but in what its story and survival reveal about history, objects and memory.

The tales and tails of Aesop’s fables

Take the unknowns, for instance. “Things where you have to fill in the blanks are more alluring,” insists Nick Cullinan, director of the British Museum. “The fact that there’s something missing invites you in…to engage with it and think about why it was made, what’s missing, and what that means.” Think of everyone who flocks to the armless “Venus de Milo” in the Louvre. Was she once carrying an apple, or modestly trying to reach so the drapery would not slip below her hips? The absent limbs keep people guessing.

There is also the question of why it is even called a tapestry at all when it is really an embroidery (the former requiring a loom, the latter done by hand). But that is a mystery only for the uninitiated: a French scholar in the 18th century called it a tapisserie (tapestry), and it stuck. “It was a bloke not knowing the difference between his weaving and his embroidery,” says Leonie Hicks, a professor at Canterbury Christ Church University.

Quite the yarn

Study the tapestry long enough and it will make you consider deeper questions, too, such as who determines people’s understanding of the past. (This is an idea pored over by plenty of historians and the wisest of sources, “Hamilton”: one song in the musical is called “Who Lives, Who Dies, Who Tells Your Story?”) Maybe a decade, and possibly more, elapsed between the end of the Battle of Hastings and when the tapestry was made: how much rumour and hearsay are woven into it as a result?

And then there were those in the 19th century, armed with needles and thread, who set about “fixing” it. Everyone who has heard of the year 1066 knows that Harold was eighty-sixed by an arrow to the eye. But was he? The earliest studies of the tapestry do not show that an arrow was involved. What you see on the tapestry is the work of a restorer, who may have taken creative liberties. Michael Lewis, a lead curator of the British Museum’s exhibition and co-author of “The Story of the Bayeux Tapestry”, thinks the arrow was an “accidental addition” by restorers in the 19th century, who were basing their work on a drawing from the 1720s that had included the arrow.

Harold gets the sharp end of the stick

Further embellishments may exist elsewhere, especially towards the end. “The arrow to the eye is probably the most famous and important one,” but there are others reconstructed in “an imaginative way”, says Dr Lewis, including a man wrapped up in vines. A lesson of the Bayeux tapestry, then, is that history is written not just by the victors, but by the restorers.

The work of an imaginative restorer

For when these fixers set about their work in the 19th century, much had been lost to them already. The historical record is funny like that: how grains of humour and information are worn down by tides of time. The result is that some scenes are now impossible to decode. The most perplexing one features the only named woman, Aelfgyva, in a doorway, speaking to a cleric who is either caressing or striking her. Nearby a naked man squats, mirroring the priest’s gesture.

One of the tapestry’s most alluring mysteries

It may have been a sexual scandal so familiar that it required no explanation. It makes you wonder which modern inside jokes and references will still be understood 900 years from now. Will anyone remember Monica Lewinsky?

One of the most notable things about the tapestry, beyond the sheer unlikelihood of its survival—including almost being used as wagon covers during the French revolution—is that so many leaders have esteemed it for different reasons. Napoleon displayed it in Paris at the Musée Napoléon (now known as the Louvre) to drum up public support for his campaigns and encourage people to enlist. The Nazis eyed it as a chronicle of Aryan supremacy and even stole small fragments of it, which Germany returned this year.

Now it is the turn of Emmanuel Macron, the French president, to use the tapestry for a different goal: to strengthen Anglo-French ties. The hard-fought loan to the British Museum seemed a long shot. Five previous attempts to borrow and exhibit it in London (including on the 900th anniversary of the conquest) resulted in British defeat. But Mr Macron triumphed where others failed, overcoming the concerns of some French curators that shipping the delicate tapestry would lead to damage. It helped that the Bayeux Museum, its usual home, was embarking on a long-overdue renovation.

On September 2nd Mr Macron and King Charles III met at the British Museum to view the tapestry and celebrate the loan. Champagne was noticeably not on offer, but English sparkling wine flowed. So did the jokes. It could have been called “the Canterbury tapestry”, King Charles mused. Mr Macron noted the tapestry’s portrayal of wine being loaded onto Norman ships bound for battle in England: “the French always having certain priorities”. It was a celebratory evening of good lines and great moods.

Macron and King Charles, battle buddies © Getty Images

The fragile tapestry looks rather stronger than politicians in this day and age. Sir Keir Starmer, who negotiated the loan, is now out. The year “1066 was not the only time we had a rapid succession of leaders,” quipped George Osborne, a British former politician who is now chair of the British Museum. Mr Macron will leave office next year and is sure to see the loan as one of his signature achievements.

Sew far, sew good

In short, national borders, allegiances and hairstyles may change over the centuries, but objects are still important in diplomacy. Later this year, in exchange for borrowing the tapestry, several works from the British Museum’s collection will travel to France. Though the French asked for the Rosetta Stone, they had to settle for other treasures, such as medieval chess pieces.

As for how to play the game to secure a coveted ticket to the exhibition, many are going to lose. The tapestry is being laid flat and stretched out in a straight line: you have to hover over a clear case to see it. This limits the number of people who can go through, as do French rules on how much light it can be exposed to in a single day. “We could sell this exhibition out many, many times over,” says Mr Cullinan of the British Museum. “We want as many people to see it as possible, but we want everyone who comes to have a really positive experience.”

For those who miss out in London, they can travel to Bayeux once the museum reopens and see the celebrated tapestry there. Many Brits are surely already plotting their Norman invasion. ■

Illustrations: Phil Hackett
Tapestry photographs: Antoine Cazin/La Fabrique de Patrimoines en Normandie

ch125.xhtml

Museums and galleries are turning to individual patrons

As corporate sponsors retreat, rich aesthetes rush in

Photograph: Jonathan Yeo Studio 2026

Sep 3rd 2026

One name gets top billing at the Bayeux tapestry show at the British Museum, and it’s not William the Conqueror. It is Igor Tulchinsky, the boss of WorldQuant, an asset-management firm. He is the only listed sponsor, so the blockbuster show is “presented by” him. He reportedly stumped up £5m ($6.8m).

In Britain individual patrons are in. At museums and art galleries, their names are increasingly visible. Whereas high culture in continental Europe is still largely state-funded, Britain is moving closer to the American model of private support for creators and curators.

Patrons’ rise partly reflects a squeeze on public funding. Tate, a group of galleries, received £56m in 2009-10 and £54m in 2024-25: a fall of about a third in real terms. The British government urges museums to foster a “culture of giving”, like their American peers. Museums—particularly those with free admission—have had to up their fundraising game.

Individual donors are also, to some extent, replacing corporate sponsors. Research by Arts and Business, a charity, found that corporate backing for culture in Britain fell for four straight years after 2007, when the global financial crisis began. And as anti-corporate protests grow bolder, such sponsorship is more likely to bring bad publicity for donor and recipient alike. BP, an oil firm, was once a prominent sponsor of culture in Britain. But museums and galleries, being open to the public, are soft targets for activists, who have tarred and feathered the entrance to the Tate (using a non-toxic tar substitute) and posted a spoof exhibit at the British Museum on plans to drill for oil under Stonehenge.

All of which has made private patrons more appealing. Big institutions have brought in specialists to woo them. In 2024-25, 15 national museums in Britain brought in £1.2bn through fundraising—three times their haul in 2023-24. Most are screening donors more carefully, too. The National Gallery has a committee to investigate “ethical and reputational issues” associated with gifts.

Patrons enjoy all kinds of benefits. Their donations can sometimes attract tax breaks. Being associated with a prestigious museum confers status, which some rich people apparently crave. And many donors simply love the arts and are happy to bankroll them.

Many are eager to support specific causes they hold dear, says Clarissa Levi, an art lawyer. Mr Tulchinsky has said that the “symmetries and proportions” in the Bayeux tapestry appeal to his quantitative sensibilities. However, he was impelled to support the exhibition by his belief in the importance of education, shaped by his own experience of moving to America from Soviet Belarus as a child. His donation will pay for learning materials for schools in the hope that “bringing young people face-to-face with this extraordinary artefact will create a lasting impact.”

There can be drawbacks to relying on philanthropists. Being human, they may be fickle or bossy. The National Gallery says it relies on private support to help fund “several senior curatorial posts”, among other things. Pushy billionaire donors might withdraw support for curators they do not like or try to have shows they disagree with axed.

The history of art is full of overweening patrons. Michelangelo had to deal with one who, being pope, thought he was infallible, which was annoying. But the great advantage of rich individuals is that, by contrast with governments, there are plenty to choose from. ■

ch126.xhtml

With “Agrippa” Robert Harris is back to roaming in ancient Rome

His new book is timely but less of Agrippa, as you might say, than previous works

Photograph: Alamy

Sep 3rd 2026

Agrippa. By Robert Harris. Doubleday; 384 pages; $32. Hutchinson Heinemann; £22

HOW OFTEN, runs a popular meme, do men think about the Roman Empire? For the British novelist Robert Harris, the answer would seem to be: every few years. Mr Harris’s first foray into the Roman period, with “Pompeii” (2003), came about because he wanted to write a dystopian novel about America. After stumbling upon a newspaper article about excavations at Pompeii, he decided that the Roman period offered a better way to satirise the complacency of a decadent superpower. He later wrote a bestselling trilogy of novels based on the life of Cicero, the last of which was published in 2015, charting the erosion of democratic norms in the dying days of the Roman Republic.

His historical thrillers consider past eras that mirror present-day concerns. “Act of Oblivion” (2022), about the national self-harm of the English civil war, has strong Brexit overtones; “An Officer and a Spy” (2013), about the Dreyfus affair, has parallels with internet-era surveillance, paranoia and whistleblowing. His thrillers set in the present day have led to popular screen adaptations, notably “The Ghost Writer” (2010) and “Conclave” (2024).

After more than a decade, Mr Harris has now returned to ancient Rome in a new novel, written from the perspective of Marcus Vipsanius Agrippa, the lifelong friend and general of Octavian (the future emperor, Augustus). The low-born Agrippa and aristocratic Octavian were boyhood friends and, as in many of Mr Harris’s novels, his narrator’s outsider perspective helps the reader navigate what may be an unfamiliar historical milieu.

It is a time of chaos, with the Roman world riven by internal conflict. Power and loyalty matter above all, and people will do anything to remain in the supreme leader’s good graces, even if that means abandoning their principles or trampling long-standing rules. Political appointments depend on family connections and marital ties. To impose peace and order, the emperor claims he must centralise power.

The modern parallels hardly need to be pointed out, which may explain the current resurgence of interest in Roman history. But few people want to wade through all six volumes of Edward Gibbon. Hence the vogue for repackaging Roman history for mainstream consumption—from the spectacle of “Gladiator 2” to the rebranding of Marcus Aurelius’s “Meditations” as a self-help book.

“Agrippa” is the latest example. As wars and intrigues rage, readers follow the ebb and flow of Agrippa’s friendship and unspoken rivalry with Octavian. Initially he rarely questions whether staying close to power and helping his friend pick off his enemies are the right things to do. Yet eventually he wonders whether his loyalty to Octavian has cost him his honour.

For much of the book Mr Harris seems hamstrung by his choice to frame his story as an end-of-life memoir written by Agrippa, depicted as a no-nonsense soldier who disdains politics and is happiest on a battlefield. In an afterword, Mr Harris says the inspiration for his characterisation was the matter-of-fact inscription on Agrippa’s most prominent surviving legacy, the Pantheon in Rome, which essentially says “Agrippa made this”. Not a man for literary flourishes, in short.

As a result “Agrippa” lacks the wit and wordplay of Mr Harris’s Cicero novels. But it gets going in its final third, as Octavian becomes Emperor Augustus. By this time Agrippa has gained some political savvy, but is also inescapably bound to Augustus by the ruthless acts they have committed to maintain his grip on power. Agrippa starts to consider how the world might look if he had made different choices—a question heightened by a clever plot twist.

Ultimately his choice of narrator means this swords-and-sandals tale does not quite result in the philosophical meditation on friendship, loyalty and power that Mr Harris aimed for. But the author’s many fans will welcome this return to the Harris Ciceronic Universe. And for everyone else, Mr Harris’s novels provide an enjoyable primer on the grandeur that was Rome. ■

ch127.xhtml

What do a spoiled Venezuelan and poor Texan have in common?

Villainy, facilitated by enablers who helped them thrive, as two new books show

Photograph: Getty Images/Pinellas County Sheriff’s Office

Sep 3rd 2026

Catch the Devil. By Pamela Colloff. Knopf; 320 pages; $32. Endeavour; £22

The Jackal. By Joby Warrick. Scribner; 384 pages; $30. Swift Press; £22

BIOGRAPHICALLY, the two men could not be more different. One is a worldly Venezuelan polyglot, the son of a champagne communist who named him and his brothers after a “hero” of the Soviet revolution. The other was the son of a violent refinery worker; he grew up in a shabby one-bedroom house in south-eastern Texas and volunteered for Lyndon Johnson’s presidential campaign. The former became the most notorious terrorist of the 20th century and is living out the end of his days in a prison outside Paris. The other cycled through jails, wives and multiple personae before dying, overweight and alone, in a nursing home near Dallas in 2020.

Both are the subjects of new biographies that illuminate the systems and circumstances that help villains thrive. The terrorist, Ilich Ramírez Sánchez (pictured, left), was a Marxist-Leninist and early practitioner of terrorist spectacle, motivated by leftist politics and self-enrichment. After lobbing a grenade into a crowd in Paris in 1974, leaving two dead and 34 wounded, then killing two policemen, he acquired a sobriquet—Carlos the Jackal—from the first name used in a false passport and a spy novel found in an apartment where he had stayed.

He was behind several headline-grabbing attacks in the 1970s and 80s, which included occupying OPEC’s headquarters in Vienna, launching rockets at Israeli planes, bombing buildings and trains in France with the aim of freeing his future wife from prison and bombing Radio Free Europe’s headquarters in Munich. In 1994 the CIA finally captured him in Khartoum and handed him over to France, where he is currently serving three life sentences.

Mr Sánchez is a type probably familiar to anyone who attended university: the spoiled rich boy who embraces leftist politics to make himself a hero. He championed Marxism, Palestinian liberation and, later, “revolutionary Islam” while indulging his tastes for luxury, booze, women and violence. One former confrère noted he took “pure pleasure in killing”. He boasted to Joby Warrick, the book’s author and a veteran national-security reporter for the Washington Post, “I killed 83 people…Another thousand were killed on my orders.”

Why did he become a murderous terrorist rather than just a petty criminal or shouting serial demonstrator with a few ill-advised piercings? Because he found his way to the Popular Front for the Liberation of Palestine (PFLP), a Marxist group, just as its co-founder, Wadie Haddad, embraced terrorist spectacle as a tactic. “We must become…a bacillus that gets under the skin of the developed world,” Haddad told a subordinate. That attitude let Sánchez play revolutionary hero while indulging his bloodlust on a global stage.

The protagonist of Pamela Colloff’s “Catch the Devil” also benefited from enablers—not other outlaws, as with the Jackal, but those operating within the legal system. Paul Skalnik (pictured, right) was charged with grand theft, larceny, forgery, sexual assault of a child and more. He served far less time than he should have because he made himself useful to police and prosecutors: other inmates seemed to have an uncontrollable urge to confess their crimes to him in private. Snitching on fellow jailbirds can reduce an inmate’s sentence, which is why police and prosecutors should treat such revelations with caution.

Skalnik lived in Pinellas County, Florida, at a time when the chief prosecutor pursued what Ms Colloff calls “a law-and-order agenda whose mandate was to put away as many people as possible for as long as possible”. His testimony won several people execution sentences, including Jim Dailey, a hometown hero-turned-alcoholic after serving in Vietnam, who appears in Ms Colloff’s account to be innocent but remains on Florida’s death row at the age of 80.

Both Skalnik and Mr Sánchez benefited from systems that rewarded their deepest, worst impulses. Maybe each would have found their way to mischief anyway. But readers cannot help wonder several things. What if police and prosecutors had been just a little more sceptical of a man who seemed to provide perfect information on demand? And what if the PFLP had sent Carlos packing? Ms Colloff’s tale is more an indictment of a system than a pure biography, as Mr Warrick’s is. But both books are terrific tales: difficult not to devour in a single setting, and difficult not to wrestle with for days afterwards. ■

ch128.xhtml

Hyrox, the fitness craze that runs and runs

A gruelling mix of running and strength-training has people working up a sweat

Photograph: Hyrox

Sep 1st 2026

INside Strawberry Arena, a cavernous stadium in Stockholm, thousands of chiselled men and women are puffing and slick with sweat. A few are the colour of the fruit. And little wonder: they have run a kilometre, covered the same distance on a ski-training machine, then run again. Between laps of the stadium they have also rowed, lunged, done burpees, carried kettlebells, pushed and pulled a sled and lobbed a medicine ball. A respectable race time is about 90 minutes; the fittest strive to finish in under an hour.

This cross between a relay race and high-intensity interval training is called Hyrox. (The name is believed to be a portmanteau of “hybrid”—a reference to its mix of cardio and strength exercises—and “rock star”, to make it seem cool.) To the unathletic, it may all sound nauseating. But Hyrox claims to be the “fastest-growing mass-participation fitness concept in the world”. At the first race, in Hamburg in 2017, some 650 people took part. In the 2026-27 season more than 2m will compete across six continents. September alone features events in 12 cities, including Athens, Mumbai and Washington.

The format has certainly set pulses racing. The hashtag #hyrox has been viewed more than 500m times on Instagram and TikTok. Peloton, a home-exercise firm, offers Hyrox training classes. Later this year a four-day European cruise will allow athletes to train at sea, with tickets for a cabin starting at a bracing £2,200 ($3,000). There is a Hyrox dating app and bank card. “Hyrox is not just a race or an event,” one acolyte proclaims online. “Hyrox is a lifestyle.”

The hype is partly explained by a boom in gym culture: men and women are pumping iron more than they used to. Across America, Britain and Europe, fitness-club memberships have reached a record high. Isabella Helm, a 25-year-old trainee accountant, loves Hyrox and trains six days a week in London: “You get addicted to feeling fitter and stronger.”

Photograph: Hyrox

Hyrox adds a social element to what is generally a solitary pursuit. There is also the adrenalin rush that comes with competition. Leaderboards foster friendly rivalry. Athletes can race solo, in pairs or teams; friends can pay to watch on in comfort from the stands.

Participants often turn races into mini-holidays. As the format spreads across the globe, there is no shortage of new destinations to visit. Danny and Zoe Rae are British influencers who, like a number of couples, combine races with recreation and doing “something nice”. “It’s a great way to travel,” Mr Rae avers.

The sport attracts racers of all ages. The average age of participants is 36, but the oldest at the world championships in Stockholm was 75. Laura Hitt trains several older people at her gym in Exeter, in south-west England. She says that “it’s one of the most accessible sports I’ve been part of.” Hyrox is kinder on the body than other fitness races—such as CrossFit, which peaked in popularity in the 2010s—because it does not require complex weightlifting. This may explain why it appeals to retired athletes as well as hobbyists.

Will Hyrox run out of breath? Some grumble that tickets are scarce and cost upwards of £150—even without accounting for travel, hotels and months of gym fees. But Moritz Fürste, the co-founder of Hyrox (and an Olympic-medal-winning hockey player), believes the format has room to grow, and hopes it will be included in the Olympics in 2032. He concedes that it is expensive, but that demand is still far exceeding supply. “It’s like tennis and golf,” he observes: “Once it becomes your sport, you’ll keep coming back.” ■

ch129.xhtml

Guo Lanying was the singer everyone loved

The performer and champion of traditional opera died on August 11th, aged 97

Photograph: Alamy

Sep 3rd 2026

The question was simple, but it still took her by surprise. A well-dressed young man stood up at one of her talks and asked, “Do you have a signature song?” Before she could answer, the whole audience, young and old, burst out in chorus. “A great river flows, wide are its waves,/Wind blows the scent of rice blossoms to both shores…” And Guo Lanying rocked with laughter.

That song, “My Motherland”, was indeed her signature. Both melody and words were in the simple folksong style she loved. She had sung it most famously in “Shangganling”, a film of 1956 about a recent battle of the Korean war, sitting demure and radiant in military uniform in a cave full of battered soldiers who, group by group, joined in with her. China was indeed beautiful, its roads broad and its waters full of white sails. But “In order to usher in a new era/They’ve woken the sleeping mountains/And changed the shapes of the rivers.” Revolutionary progress was mixed in with natural human homesickness. Another of her most famous songs, “Nanniwan”, celebrated the transformation of a north-western gorge into fertile, productive fields (“Crops everywhere!”) by the 359th Brigade. Again, people sang this up and down the land, and she was delighted to perform it to anyone, many or few, farmers or factory workers, premiers (Zhou Enlai was a great fan) or dumpling-sellers. Not for nothing was she named in 2019 a People’s Artist, one of China’s highest honours.

Yet it was opera that had made her: traditional opera, dealing not just with lords and princes (she parading in heavy, glittering robes and extravagant headresses) but, from the 1940s, with ordinary people trying to get by. At 15, she fell in love with an opera called “The White-Haired Girl”. This was the story of Xi’er, a young peasant woman whose land was taken by a vicious landlord who made her his concubine. Rather than submit she took refuge in a cave, her hair and skin turning white as a ghost’s, her clothes in rags, until the soldiers of the Chinese Communist Party brought her back to the sunlight.

Lanying thrilled to this role, in large part, because a rapacious man got his come-uppance. In China’s old society, even sometimes under the new, women were on the bottom rung: subservient to husbands and parents-in-law, confined to the house. She sang their resistance. In another opera, “The Injustice to Dou E”, she played a woman falsely accused of murder, singing much of the time on her knees with a wooden crime placard strapped to her back. But her final aria, “Pathetic and Disgraceful”, was a lacerating attack on her male abusers and accusers and on the old justice system in general. The high, pure soprano could become a furious mezzo, as necessary.

That voice had been beaten into her, as was the way in those days. As an apprentice singer and performer, on a five-year contract starting at the age of six, she was taken into her master’s household and ruthlessly trained. If she failed to do a somersault perfectly, she was lashed with a leather sheep-whip. For insubordination, she was bludgeoned black and blue with a rolling pin. In the bitter winters of Shanxi, her home region in the north, she was made to do vocalising exercises lying face-down on a frozen lake; she had not done well unless her breath melted a hole in the ice.

To escape this regime she once took poison and, at 11, ran away, but being so far from her own village she had no idea where to go. Besides, she was determined to sing, even if this treatment was the price. She had been spared to do it. As a baby, unable to get a drop of mother’s milk because her peasant parents had only a handful of millet in an iron pot, she had been left in the woods to die. But an aunt rescued her, and an uncle, when she was four, suggested opera. Her parents, starving again by then, sold her to the opera master for five copper coins, or around $12. She fought to stay at home but her mother remarked that, if she became a singer, she could eat from a hundred households.

So she had persevered, progressing from maids-with-teapots parts to roles big enough to get her on the posters. But after she had seen “The White-Haired Girl” she packed her few things and ran after the performance troupe, joining the revolution by begging for the part of Xi’er. In order to get it she practised harder than ever, especially enunciation. Since audiences listened to opera not for her voice, but for her meaning, she must bite on every word, and let each one land like a hammer-blow on the heart.

Yet, as an art form, Chinese opera was struggling. During the Cultural Revolution it was banned for a decade and she herself was sent to “the cowshed”, an improvised prison where she had to do forced labour. (In December 1976, after the arrest of the Gang of Four, she dared to change a song in mid-concert to praise, not the People’s Liberation Army as usual, but Zhou, whose death that year had hastened the end of the catastrophe.) Zhou had hinted to her that she ought to train successors in her art before it was eclipsed by Western influences. So, on retirement in the 1980s, she sold her house and her awards to found in Guangzhou an art and opera school that was later named after her. She and her husband even helped build it, heaving stones and laying roads.

Becoming China’s favourite female singer had been equally hard work. She impressed that work ethic on her students; if she wasn’t tough on them, audiences would be tough later. Her routine in old age was still vigorously ascetic, including half a bowl of aged Shanxi vinegar with her meals. She painted orchids (lan in Chinese), feeling empathy with them from the name she shared. She had endured. It was true that a deep core of anger had also pushed her on. But to dissolve it, and forget it, she had only to hear a packed, adoring theatre singing her songs back to her. ■