Wall Street’s artificial intelligence boom is facing its biggest test yet.
Nvidia plunged 14% on Friday, the Nasdaq Composite lost 5.3% and the entire S&P 500 entered correction territory as investors questioned whether the trillions of dollars invested in AI infrastructure would ever produce the expected returns.
The Nasdaq closed at 22,617.73, about 18% below its October 6 closing record of 27,599.79. Its 10% weekly decline was the worst since 2020, as the index has now erased its 2026 gains.
Now you can act on the news you read: go long or short Nvidia with perpetual futures, all from your Decrypt Money account.
Nvidia, the company behind much of the AI rally, ended Friday at $125.53, down 47.5% from its October closing record of $239.24.
The chipmaker lost nearly 23% during the week as fallout from former CEO Jensen Huang’s departure and reports of an SEC accounting investigation added to concerns about the AI business model.
Other AI-related companies suffered even greater losses. CoreWeave, which rents computing capacity to AI developers, closed at $39.51, nearly 80% below its 2025 high, while Oracle fell about 70% from its highs and ended the week at $101.19.
AMD lost 20.5% for the week, closing at $424.39, while Alphabet lost 10.3% to $265.29. The iShares Semiconductor ETF, SOXX, fell 17.4% during the same period, showing that investors are selling all companies supplying the AI industry, not just its more speculative businesses.
The sell-off accelerated just as Anthropic completed its historic $81 billion IPO. Promoter Claude priced his shares at $83 and closed his first trading session up 17%, but his shares fell another 9% on Friday as enthusiasm over the offering failed to restore confidence across the industry.
With so many stocks in the red, it makes sense that many AI fans are starting to get nervous.
Polymarket traders have become increasingly pessimistic. Contracts assessing whether the AI industry will experience a defined downturn had a probability of approximately 54% as of December 31, 2026 and 65% as of June 30, 2027, according to Friday’s market overview.
The market attracted about $4.1 million in trading volume, but its definition of a bubble burst is more demanding than that of a falling stock market. At least three specific events must occur in a single 90-day period, including Nvidia closing 50% below its all-time high, SOXX falling 40%, major AI vendors experiencing similar collapses or serious financial problems among AI developers.
SOXX has already fallen about 43% from its split-adjusted record closing price, meeting one of the price thresholds specified on paper. Nvidia is getting close, even if crossing either threshold would not be enough to settle the contract.
Bitcoin and crypto are dragged into the sell-off
The damage also extended to cryptocurrencies, undermining expectations that 2026 would bring a lasting recovery. Bitcoin ended the week near $58,895, down 9.8% over the period and more than 50% below its 2025 peak.
The total market capitalization of cryptocurrencies has fallen to around $2.01 trillion, from over $4 trillion at its peak. As Decrypt reported on Friday, the Crypto Fear and Greed Index fell to 9, signaling extreme fear and its lowest reading since February.
It was supposed to be the year of the usual moon before the halving. Instead, for the first time in Bitcoin history, the coin is about to experience two red years in a row.
Bitcoin increasingly behaves like other speculative assets sensitive to liquidity conditions, particularly when investors expect borrowing costs to remain high.
The pressure is also visible in investment flows. U.S. spot Bitcoin ETFs recorded $582.2 million in net withdrawals on Thursday, while forced liquidations added to selling as Bitcoin fell below the psychologically important $60,000 level.
Is this a dot-com 2.0? The experts have something to say
Owen Lamont, a portfolio manager at Acadian Asset Management and a former finance professor at the University of Chicago, says the warning signs have become hard to ignore. After claiming in January that AI stocks were not yet in a bubble, he said Fortune this week, he now assigns a 70% probability that the U.S. stock market is already in one.
Lamont tracks four warning signs: excessive valuations, widespread belief in a speculative boom, new stock issuance, and investors pouring money into the market. Its latest assessment suggests all four are now present, although its emissions figures are heavily influenced by SpaceX’s massive IPO earlier this year.
Bank of America’s survey of fund managers in November offers another indication of deteriorating sentiment. Investors have reduced their technology exposure the most since 2006, and 52% expect at least one major AI hyperscaler to announce a reduction in capital spending within six months, Fortune reported.
The European Central Bank has also raised concerns about how AI development is financed. Its November report estimates that capital spending by large U.S. technology companies could reach $5.5 trillion cumulatively by 2030, with increasing reliance on debt as infrastructure spending outpaces internally generated cash.
Nvidia shares fall to 52-week low after SEC investigates accounting
The similarities to the Internet era are hard to ignore: Investors assign huge valuations to companies based on expected demand, companies rush to build infrastructure before they have demonstrated profitability, and financial markets are increasingly dependent on a small group of technology companies.
During the initial collapse of the dot-com bubble, the Nasdaq lost about 78 percent between March 2000 and October 2002. The current decline is considerably less, but history also reminds us that a technology can succeed commercially while the companies financing its expansion lose enormous amounts of market value.
There are, however, important differences. Many of today’s biggest beneficiaries of AI, including Nvidia, Alphabet and Microsoft, operate profitable businesses with substantial existing revenues, unlike many internet startups whose valuations depended almost entirely on their future growth.
Nvidia’s latest quarterly results illustrate the contradiction. The company beat earnings estimates, reported record revenue and net income, and issued higher-than-expected revenue guidance, but its shares continued to slump.
Richard Saperstein, chief investment officer at Treasury Partners, described the economic downturn as a “repricing of AI” in comments to Reutersciting slowing data center construction and tightening financing conditions. Lawrence Gillum, chief fixed income strategist at LPL Financial, offered a less alarming interpretation, saying the decline was “not a bubble burst” and arguing that the AI investment cycle still has room to run.
Higher interest rates complicate both arguments. Investors increasingly expect the Federal Reserve to raise rates again in December, which would raise data center financing costs while reducing the appeal of stocks whose valuations depend on earnings in coming years.
The totality of the evidence suggests more than a simple nervous breakdown, but less than a confirmed systemic collapse. Stocks have revalued faster than the underlying technology has changed.
Nvidia Board Halts Return of Jensen Huang, Starts Search for New CEO
What happens next will depend on whether real demand for AI can justify the data center spending spree and whether the companies funding it can resist tighter credit. The real test isn’t just whether Nvidia can continue selling chips. It’s a question of whether its customers can turn all that computing power into profits.
Right now, much of the hype about AI is just…hype. Markets evolve more based on the perceived potential of AI in the future and less based on what AI actually does now.
Crypto investors have already seen this with ICOs, TPS wars, the metaverse, NFT fever and more. For many of them, this distinction brings little comfort.
The Bitcoin crisis shows how quickly markets can stop distinguishing between promising technologies and overvalued bets when trust disappears. The Internet crash did not kill the Internet; it killed the assumption that every Internet company would make its investors rich.
AI could be the next technology that changes the world and proves that being right about the future is not the same as making money from it.