美国股市崩盘被广泛预测，因美股科技估值处于高位，由“壮丽七巨头”和对人工智能的押注推动。标准普尔的周期性调整市盈率已回到互联网泡沫时代的水平。摩根大通估算，为了在2030年前对预期的人工智能资本支出获得10%的回报，公司们每年需实现6500亿美元的人工智能收入——相当于每位iPhone用户每年超过400美元——使得这些预期在数字上令人却步。与2007–09年不同，今日的繁荣主要由股本融资推动。

脆弱性集中在美国消费者：股票现在占家庭财富的21%——比互联网泡沫高峰时约多四分之一——与人工智能相关的资产占去年美国财富增长的近一半。与互联网泡沫相当的股市暴跌将把美国家庭净值削减约8%，我们估算这将使消费者支出回缩约相当于1.6%的GDP，并可能把本已走弱的劳动市场推入衰退。外国投资者持有价值18万亿美元的美股；今年美元已下跌8%。

溢出效应将冲击低增长的欧洲和通缩风险的中国，加剧出口商的困境并扩大中国制造业产能过剩。政策制定者有应对手段：美联储有空间降息，一些国家可采取财政刺激，但发达国家债务总额约为GDP的110%，赤字将扩大，债市压力可能上升，政治和保护主义紧张局势会加剧，即使衰退相对浅薄。

An American stockmarket crash is widely predicted because US tech valuations are stratospheric, driven by the “magnificent seven” and AI hopes. The S&P’s cyclically adjusted price-earnings ratio is back at dotcom-era levels. JPMorgan estimates companies need $650bn of annual AI revenues by 2030 to earn 10% on projected AI capex—over $400 per iPhone user—making expectations numerically daunting. Unlike 2007–09, today’s boom is mostly equity-financed.

The vulnerability centers on the American consumer: stocks now equal 21% of household wealth—about a quarter more than at the dotcom peak—and AI-related assets account for nearly half of last year’s wealth gain. A dotcom-sized stock fall would cut US household net worth by about 8%, which we estimate would reduce consumer spending by roughly 1.6% of GDP and likely tip an already weakening labour market into recession. Foreign investors hold $18trn of US equities; the dollar is down 8% this year.

Spillovers would hit low-growth Europe and deflationary China, worsening exporters’ troubles and enlarging China’s manufacturing glut. Policymakers can respond: the Fed has room to cut rates and some countries can use fiscal stimulus, but global public debt equals roughly 110% of GDP, so deficits would widen, bond-market stress could rise, and political and protectionist strains would intensify even if the downturn is shallow.