Citadel's Ken Griffin Says Investors Are 'Almost Certainly' in an AI Bubble, Nvidia Closes Down 0.29%

Griffin said in May that he was depressed by what AI could achieve and called it "garbage" at Davos earlier this year, and on Tuesday, Oct. 20, he landed on a new word: bubble.

Speaking to Bloomberg's Dani Burger at Citadel's Future of Global Markets Conference in New York, Griffin said investors are "almost certainly" in an AI bubble. "Any radical transformation of our economy is generally intertwined with a bubble," he said.

His sharper point was about customers, not valuations. "Corporate America has been very uneven in its ability to extract value from its AI models," Griffin said. "I think all of Corporate America is looking at their bills from Anthropic and OpenAI and going, 'Okay, we need to think about how to make sure we're getting productivity out of this.'"

The debt behind the build-out

The companies leading the build-out are Amazon, CoreWeave, Google, Meta, Microsoft and Oracle. They had issued $193 billion in bonds as of early August, up from $16.7 billion in 2024, according to LSEG data. Separately, iShares counts nearly $200 billion of investment-grade debt from Amazon, Microsoft, Alphabet, Meta and Oracle in the first half of 2026, almost double all of 2025.

The Brookings Institution projects that spending on data centers, networking and related equipment will total $10.3 trillion from last year through 2032.

Stocks have softened as investors have focused on the debt. The Nasdaq composite fell more than 100 points over the past two weeks, and the S&P 500 slipped roughly 90 points.

Griffin's view of recent chip stock losses

Griffin also addressed last week's chip stock sell-off. Everyone blames hedge fund "deleveraging," he said, and "the reason that people love that answer is it actually washes your hands of accountability."

He did not name what he thinks the real cause was. His comments about uneven returns on AI spending point toward the question of whether the payoff is arriving on schedule.

Not a bearish call

Griffin is not betting on collapse. He pointed to the dot-com era and its "fanciful" business models. "The companies that are on the other side of that bubble are the market leaders for decades to come," he said. "I just wish I'd kept my Amazon stock in that moment."

He described a split effect. "You're seeing the threat to white collar jobs playing out; as a business owner, you're seeing your competitive moats be filled in by bulldozers," he said. "At the same time, you're thinking that this will be a golden era for entrepreneurs."

Griffin said Anthropic's fortunes hinge on convincing investors its models are indispensable to businesses in every sector.

Anthropic's filing and the $750 billion question

Anthropic is expected to file publicly for a stock market listing in the coming weeks. The offering could rank among the largest IPOs ever. The Wall Street Journal has reported that the company could put $750 billion toward additional computing capacity over the next several years.

That is the figure Griffin's finance teams are staring at from the other side of the ledger. Customers are paying the bills, and the filing is expected to show how much more the supply side plans to spend.

Other warnings, and the counterargument

Griffin is not alone. Earlier this month Bridgewater founder Ray Dalio told the Forbes Global CEO Conference in Singapore that AI has the hallmarks of a "classic bubble." "I think we're close to that," he said of the bursting point. Dalio named the heavy debt raised to fund AI investment as the major risk.

Economist Paul Krugman has argued the AI boom is crowding out investment elsewhere in the economy, something he said the dot-com boom did not do. "This isn't a hypothetical risk," he said. BCA Research has said total factor productivity growth, adjusted for utilization, was zero in the second quarter, meaning no economy-wide productivity acceleration yet.

There is a counterargument from people inside the industry. Visser, head of AI Macro Nexus Research at 22V Research, says a mismatch between demand and supply, not valuation, is the defining feature of a bubble. He says the demand is coming from AI agents, and he expects their number to grow from thousands today to billions next year. Nvidia's CFO has described the company's outlook as "supply-constrained."

Griffin's own firm offers an example of the upside. Citadel's agentic AI tools cut certain research work from six weeks to two hours, according to Yahoo Finance.

How the market reacted

It barely reacted. Nvidia closed Tuesday at $216.47, down 0.29%, and rose 2.12% to $221.06 in after-hours trading as of 7:59 p.m. ET. The Invesco QQQ Trust slid 0.10% to $717.63.

Nvidia carries its own exposure to Griffin's concern. Its supply commitments have climbed to $279 billion, mostly memory for its Vera Rubin systems. If customers cannot show a return and slow their orders, those commitments become a heavier burden.

The next hard data point is Anthropic's public filing. It is expected within weeks, and it should show how much of that $750 billion computing plan the company is willing to put in writing.