Has the artificial intelligence bubble burst? What was once a provocative question raised by bearish investors has become a central concern on Wall Street. On Thursday, November 5, US stocks lost $3.2 trillion in market value, according to the Financial Times. The Nasdaq 100 plunged 6.4%, Nvidia shed more than $400 billion in market capitalization, and the Philadelphia Semiconductor Index tumbled 12.3%. As this column was being written, Asian markets were opening sharply lower on Friday, with trading temporarily halted in South Korea. What began as a sell-off in technology stocks now threatens to develop into a much broader financial crisis.
A decline of this magnitude does not necessarily mean a bubble has burst. Financial markets experience violent corrections, often followed by equally sharp recoveries. The real distinction lies in whether investors have lost faith in the assumptions that supported previous valuations. In artificial intelligence, the fundamental question is whether trillions of dollars invested in semiconductors, servers and data centers will generate enough revenue and profits to deliver adequate returns on capital. The Democrats’ sweeping victory in the US midterm elections, coupled with the prospect of restrictions on data center construction, accelerated the sell-off. But AI-related stocks had already been losing momentum before the election that cost President Donald Trump’s Republican Party control of Congress.
The most troubling warning signs are now emerging beyond the stock market. Credit default swaps on CoreWeave, a major provider of AI computing infrastructure, have reached record levels, reflecting growing concerns about the sustainability of its debt. Meanwhile, major investors are questioning the $1.5 trillion valuation Anthropic is seeking in its planned initial public offering later this month. According to The Wall Street Journal, some believe a valuation closer to $1 trillion would be more realistic. OpenAI is also facing a less accommodating funding environment: its latest financing round had raised $11 billion against an initial target of $30 billion.
The decisive test will come from the technology giants themselves. Microsoft, Alphabet, Amazon and Meta account for an extraordinary share of global demand for chips, servers, electricity and data center construction. Their capital spending drives suppliers’ earnings, which in turn support elevated stock market valuations. If these companies begin postponing projects or cutting their 2027 investment budgets, that virtuous cycle could reverse. Lower spending would mean fewer orders, weaker earnings expectations and further declines in share prices. The slowdown could also spread to industrial companies, energy suppliers and the construction industry. At that point, a bursting AI bubble would no longer be a problem confined to technology investors.
So when will we know whether the bubble has actually burst? The next few trading sessions will show whether the sell-off turns into a broader flight from risk assets. The coming weeks, particularly decisions surrounding Anthropic’s IPO and financing conditions for heavily indebted AI companies, will reveal how deeply investor confidence has been shaken. But the clearest evidence will probably emerge over the next three months, as the largest technology companies report earnings and update their capital expenditure plans for 2027.
During the dot-com bubble of the early 2000s, repeated market rebounds encouraged investors to believe the worst was over, even as valuations continued to collapse. This week’s turmoil already bears the hallmarks of something more serious than ordinary profit-taking: historic market losses, international contagion, mounting credit stress and growing doubts about corporate financing.
It may still be too early to declare the end of the AI valuation boom. But history suggests that the bursting of a financial bubble is rarely recognized as it happens. By the time markets finally acknowledge what has occurred, the losses have usually already piled up.
Leave a Reply