- Nasdaq 100 swung from a gain of more than 2% to a loss of over 2% on Thursday, a reversal seen only three other times since 2009.
- The index is now more than 15% below its October peak, as AI valuation concerns, hawkish Fed signals, and deleveraging weigh on sentiment.
- Historically, similar extreme reversals have been followed by a median gain of 4.2% over the next five sessions, but further downside remains possible.
A Historic Swing
The Nasdaq 100 experienced one of its most dramatic intraday reversals in over a decade on Thursday, plunging from a gain of more than 2% to a loss of over 2%, according to Bloomberg. Such a swing has occurred only three other times since 2009: in March 2020, July 2020, and November 2025. The index closed approximately 2.2% lower, while the S&P 500 fell 2% for its sixth consecutive loss. More than 460 S&P 500 companies declined, and all megacap stocks ended in the red.
The reversal reflects a broader loss of confidence in the AI investment boom, not merely a bad day for technology stocks. Even a successful stock-market debut by Anthropic (NASDAQ:ANT)—the AI developer behind Claude—failed to sustain buying. Anthropic finished its first trading day at $97.17, up 17.1% from its $83 IPO price, after trading above $115. Investors still showed appetite for a major AI company, but that enthusiasm did not restore confidence in the wider sector.
AI Infrastructure Under Pressure
The reversal was particularly pronounced among AI infrastructure companies. Nvidia (NASDAQ:NVDA) reversed an early 4.7% gain into a 3.3% loss, while CoreWeave (NASDAQ:CRWV) swung from a 10.5% gain to a 6.1% decline. The selloff accelerated amid growing scrutiny of the industry’s borrowing boom, with investors questioning whether revenue and productivity gains can justify the enormous commitments to chips, data centers, and related infrastructure.
“Willingness to fund a leading AI laboratory is not the same as willingness to fund heavily indebted cloud-infrastructure providers,” investor Luke Lango told Bloomberg. The distinction between demand for AI products and the financing of AI infrastructure is becoming increasingly important.
Adding to the uncertainty, AP identified Jensen Huang’s sudden resignation as CEO of Nvidia as a recent source of instability for AI investors. Meanwhile, Advanced Micro Devices (NASDAQ:AMD) fell after a Hunterbrook Capital short-selling call raised China-related concerns; AMD denied wrongdoing.
Economic and Political Crosscurrents
Three connected concerns dominate the narrative: AI spending versus returns, expensive financing, and softer economic signals. The 10-year Treasury yield ended Thursday near 5.04%, despite declining as investors sought safer assets. High rates make debt-funded infrastructure projects harder to finance and increase the hurdle for future investment returns.
U.S. unemployment-benefit applications were slightly higher than expected, while a mid-Atlantic manufacturing report showed a larger-than-expected slowdown. However, jobless claims remained low by historical standards, and these reports do not, by themselves, establish a recession.
Political developments also weigh on sentiment. AP reports that Democratic gains in this month’s U.S. midterm elections increased investor concern about tighter AI regulation. This is a market expectation, not evidence that a specific new law caused Thursday’s decline.
A Historical Lens
After roughly 30 comparable reversals since 1985, the median gain over the following five sessions has been 4.2%, according to Bloomberg. However, a median return is the midpoint of historical outcomes—not a probability of a gain or a guaranteed return. The post-2009 count and the post-1985 sample cover different periods, and without the methodology, it is unclear whether their screening rules are identical.
A useful comparison is November 2025, when another AI-related selloff driven by capital-spending concerns, valuations, and changing expectations for Federal Reserve rate cuts was followed by a strong final-week rebound. That illustrates how sharp recoveries can occur within unsettled markets.
What’s Next
The central question is whether any rebound attracts sustained buying or is sold again. “Anthropic’s successful IPO did little to resolve the larger problems: AI profitability, massive spending, high interest rates, credit stress, and geopolitical risks,” Matt Maley of Miller Tabak told Bloomberg. That supports a cautious interpretation of the historical rebound statistic rather than treating it as a forecast.
Friday’s early indications remained weak. At approximately 5:49 a.m. EST, Nasdaq-100 futures were down 1.12%, though that is an earlier premarket snapshot. The MSCI World Index declined 1.7% on Thursday, and subsequent reporting described weaker Asian markets and U.S. technology futures heading into Friday.
The effects are not confined to speculative AI holdings. The S&P 500—widely held in retirement accounts—was 8.2% below its October record after Thursday’s losses, according to AP. Meanwhile, Bloomberg reported that two prepaid energy transactions totaling $3 billion, with Alphabet (NASDAQ:GOOGL) as funding recipient, were delayed amid weak investor demand and market volatility. SpaceX (NASDAQ:SPCX) is reportedly considering at least $32 billion of debt tied to a chip-financing arrangement, even as the Alphabet-linked energy deals encountered weak demand. These developments show that the AI investment cycle remains active, but access to capital is becoming more selective.
As the market digests this historic swing, investors will be watching whether the historical pattern of a short-term bounce plays out or if the AI-driven selloff has further room to run.