周五，整个美国股市宛如一座万圣节鬼屋，与不断膨胀的AI泡沫相关的科技股迅速下跌，拖累了大盘。
The entire US stock market looked like a house of Halloween horrors on Friday, driven by rapidly sinking tech stocks tied to the swelling AI bubble.
据NBC新闻报道，以科技股为主的纳斯达克100指数——苹果、英伟达、微软和亚马逊等主要AI股票所在的指数——在本文撰写时下跌了2.6%，很可能出现1万亿美元的市值蒸发。
The tech-heavy Nasdaq 100 — home of major AI stocks like Apple, Nvidia, Microsoft Corp and Amazon — was down 2.6 percent at the time of writing, well on track for a $1 trillion drop, as NBC News observed.
下跌由亚马逊推动，该公司股价在周五早盘交易开始时下跌约2%。尽管亚马逊第三季度财报显示其云计算业务继续保持增长势头，但这仍不足以抵消围绕AI泡沫的深层担忧。尤其是亚马逊2026年的资本支出激增至2200亿美元，其中很大一部分与AI相关。
Losses were pushed by Amazon, which shed about two percent when trading opened Friday morning. Though Amazon’s third quarter earnings report pointed to continued momentum in the company’s cloud computing unit, it wasn’t enough to offset deep-seated fears around the AI bubble. In particular, Amazon’s capital expenditures for 2026 ballooned to $220 billion, much of which is tied to AI.
周五的暴跌之前，谷歌母公司Alphabet的股价在周四暴跌超过12.6%——Alphabet是AI交易的重要风向标。这是该公司有史以来最糟糕的单个交易日，也是历史上任何公司股票第二大单日市值蒸发，抹去了超过5000亿美元的市值。
Friday’s rout comes after shares for Google’s parent company Alphabet — a major bellwether for the AI trade — plummeted over 12.6 percent on Thursday. It was the company’s worst single trading day ever, and the second-largest single day wipeout of any company’s stock in history, obliterating over $500 billion worth of market value.
这一走势正在加剧人们对AI泡沫破裂的担忧，因为数万亿美元仍集中在少数几家顶级AI公司手中，而这些公司大多难以将这项技术转化为具有商业可行性的技术。许多人现在在问，这种超大规模路线——即科技垄断企业建设庞大、高能耗的数据中心来满足AI需求——与以极低成本部署高效轻量模型的路线相比，是否是一种可持续的做法。
The motion is all stoking fears of a bursting AI bubble, as trillions of dollars remain tied up in just a few top AI companies, which have largely struggled to turn the tech into a commercially viable technology. Many are now asking whether this hyperscale approach — in which tech monopolies build massive, energy-intensive data centers to fuel AI demand — is a sustainable approach, compared to one which deploys efficient and lightweight models for a fraction of the cost.
至于这是否意味着AI盛宴的终结，现在判断可能还为时过早。正如前联邦消费者保护监管机构罗希特·乔普拉最近在接受《政治报》采访时所说，对许多投资者而言，到了这个地步，超大规模路线只能成功，别无选择。
As for whether this is the end of the AI feeding frenzy, it may be too early to tell. For many investors at this point, the hyperscale approach simply has to work, as former federal consumer protection regulator Rohit Chopra told Politico in a recent interview.
乔普拉解释说：“市场上很多人都觉得，我们别无选择，只能指望它成功。但到了某个时候，我们就会看到那些关于营收和增长的假设是否能够兑现。如果不能兑现，那时你就会开始看到整个经济出现相当显著的连锁反应。”
“Many in the market feel we have no choice but for it to work out,” Chopra explained. “But at some point, we will see whether those hypotheses about revenue and growth come together. And if they don’t, that’s where you start seeing a pretty significant ripple effect across the economy.”