Citadel's Ken Griffin warns of AI stock bubble

Citadel founder Ken Griffin warned on October 21, 2026, that investors are almost certainly in an AI bubble. He made the statement at the Future of Global Markets Conference in New York, an event run by his firm, Citadel Securities.

Major market indicators presented a mixed picture that day. The S&P 500 and Dow Jones Industrial Average posted gains, while the Nasdaq and key semiconductor funds declined.

Griffin's AI bubble warning and market positioning

Ken Griffin argued that radical economic transformations are generally intertwined with financial bubbles. He compared the current AI moment to the dot-com era, noting that while many ventures failed, companies like Amazon survived the bust to become dominant. His view on AI has shifted markedly during the year, from being dismissive of the hype at the Davos forum in January to expressing an unsettled perspective by May.

Griffin was not dismissing AI's potential. He warned that excess valuations and failed ventures may coexist with the emergence of lasting winners. In response to these concentration risks, Citadel has significantly adjusted its portfolio. The firm, which manages about $77 billion, cut its exposure by more than 80% using over 100 block trades worth more than $4 billion. Citadel earned gains by supplying liquidity during market panics. Its main multi-strategy fund, Wellington, gained 5.94% in July and was up 12% for the year to date.

Sector-specific developments and fund impacts

Other major investors share Griffin's bubble concerns. Bridgewater founder Ray Dalio has called AI a classic bubble. The risks of concentrated bets were starkly illustrated by the Situational Awareness fund, which held concentrated positions in AI infrastructure stocks while shorting software names. SanDisk and Micron made up more than 56% of its disclosed U.S. holdings at the end of June. When SanDisk fell nearly 47% and Micron dropped roughly 29% in July, the highly leveraged fund faced massive margin calls. It had used up to 400% leverage, topped near $45 billion in assets, and lost roughly $35 billion.

Despite the volatility warnings, select technology giants continue to report staggering AI-driven revenue growth. Their recent earnings provide concrete figures on the sector's current scale.

Elon Musk stated that SpaceX could exit 2027 with a revenue run rate as high as $1 trillion. He also claimed compute remains scarce, noting that older Nvidia chips now rent for more than their original price when new. Other companies are making large bets on future capacity. Anthropic is preparing to go public and may invest $750 billion toward extra computing capacity over the next several years. Super Micro Computer also released a preliminary update showing September-quarter revenue far ahead of its own guidance.

Griffin noted that corporate America is uneven in extracting value from AI models. The beginning of the mega-cap tech earnings season, starting with Meta, will provide critical evidence on whether massive AI spending is generating returns that support current valuations.