Ken Griffin Says AI Is ‘Almost Certainly’ in a Bubble, But Advises Investors Not to Panic

Ken Griffin, the billionaire founder and CEO of Wall Street giant Citadel, told an audience at Citadel’s Future of Global Markets Conference in New York that investors are “almost certainly” living through an artificial intelligence bubble—yet he stopped short of advising them to dump their AI holdings.

The remarks from one of the world’s most prominent hedge fund managers spotlight a central paradox facing modern financial markets. While speculative excess and soaring valuations have characterized the sector for the better part of two years, the underlying technology is simultaneously revolutionizing business operations and reshaping the global economy.

“What’s left in the wake of these bubbles, if we want to use that word, is often game changing for the world,” Griffin said during the October 20 event. “And the companies that are on the other side of that bubble are the market leaders for decades to come.”

The stakes of the current AI boom extend far beyond the portfolios of Silicon Valley venture capitalists or Wall Street day traders. The massive capital expenditure cycle required to build out AI infrastructure has underpinned broader equity markets over the past year. Because of this widespread integration, countless index funds, mutual accounts, and everyday retirement portfolios hold a meaningful slice of the tech sector’s high-flying stocks.

A market bubble traditionally occurs when asset prices climb far faster than the underlying businesses can justify through fundamental earnings, creating a dangerous disconnect until economic reality finally catches up. Citadel, operating as a massive private hedge fund, manages and invests capital on behalf of wealthy clients, endowments, and institutional investors worldwide, giving Griffin a front-row seat to these macroeconomic shifts.

Yet, according to Griffin, the mere presence of a bubble is not an apocalyptic event when viewed through a historical lens. “Any radical transformation of our economy is generally intertwined with a bubble,” he noted, drawing parallels to the dot-com boom of the late 1990s and candidly admitting that he wishes he had held onto his early Amazon shares rather than bailing out prematurely.

The companies that successfully navigate the turbulence and emerge on the other side of a market correction tend to dominate global commerce for decades, Griffin argued. However, he stopped short of predicting an imminent or catastrophic collapse.

In fact, Citadel’s own internal operations serve as a living case study of the technology’s profound utility. Griffin revealed that the firm utilizes advanced AI agents—sophisticated software capable of carrying out complex, multistep tasks independently—to perform research projects that previously consumed weeks of labor, compressing the timeline down to a mere couple of hours.

Despite this undeniable productivity boost, Griffin drew strategic parallels between the current market for graphics processing units (GPUs) and the economics of the airline industry. Both sectors share a characteristic known as a very low marginal cost to run. Once the heavy upfront infrastructure is paid for, the incremental expense of serving one additional customer is minuscule, meaning that prices can plummet violently once supply outstrips consumer demand.

To illustrate this risk, Griffin pointed back to the telecommunications boom around the year 2000. During that period, exuberant companies rushed to lay vastly more fiber-optic cable than actual global demand required, causing communication prices to collapse. Looking at today’s hardware market, he expressed skepticism that older generations of chips—those sitting at five or six years old, which currently command rental rates he described as extraordinary—will maintain that pricing power for another five or six years.

Corporate buyers are already growing increasingly discerning, Griffin observed. Businesses are sorting through the marketplace to determine which tasks genuinely justify premium, proprietary models from leading labs like OpenAI and Anthropic, and which routine workloads can easily be handled by cheaper, open-source alternatives that any enterprise can download and run independently on their own local hardware infrastructure.

Corporate America’s ability to extract genuine value from AI has been notably uneven, Griffin added. Many executives are currently studying expensive software bills from major labs and asking a fundamental question: where is the actual productivity return on investment?

Independent forecasts underscore the sheer scale of the financial commitments currently flowing through the ecosystem. Gartner has projected that worldwide AI spending will reach nearly $1.5 trillion for 2025 alone, encompassing everything from advanced silicon chips and enterprise servers to consumer devices equipped with localized AI capabilities. Meanwhile, Nvidia CEO Jensen Huang asserted during remarks in Davos that the technology sector is still only “a few hundred billion dollars” into what he characterized as the largest infrastructure buildout in human history, suggesting that trillions more in capital investments remain ahead.

If corporate customers eventually balk at premium subscription prices or hardware costs, the foundational chip manufacturers and cloud computing giants relying on that continuous spending would be the first to feel the financial shockwaves.

The high-stakes nature of these leveraged bets was underscored earlier in the year. On July 30, Situational Awareness—a specialized hedge fund founded by former OpenAI researcher Leopold Aschenbrenner—sold the vast majority of its public stock portfolio to Citadel at roughly a 10% discount. The distress sale occurred after the fund’s public portfolio reportedly plunged about 67% over the course of July.

The fund had aggressively utilized leverage, borrowing as much as four times its underlying capital to accumulate high-growth AI stocks. As market prices retreated, its institutional lenders repeatedly issued margin calls, demanding more collateral on a daily basis. Forced selling of this nature tends to be self-reinforcing, because each subsequent block sale drives market prices lower, which in turn triggers fresh collateral demands from frightened lenders.

Citadel subsequently stepped in, and by August, Griffin wrote in a client letter that the firm had successfully shed more than 80% of the aggregate risk inherited from that distressed portfolio. The maneuver demonstrated that even as Griffin warns of speculative excess, sophisticated market makers are well-positioned to capitalize on the fallout of forced liquidation.

Griffin’s macroeconomic commentary extends beyond technology markets. Earlier this year, he warned that if geopolitical tensions kept the Strait of Hormuz—a vital oil shipping lane—closed for a prolonged period of six to 12 months, global energy prices would rise materially and threaten to push the broader world economy into a recession.

Meanwhile, Wall Street analysts are already beginning to price in a cooling phase for the sector’s breakneck growth. According to research published by Goldman Sachs, capital spending growth—the massive sums poured into advanced chips and sprawling data centers by the largest infrastructure builders—is expected to slow from a staggering 75% year-over-year rate in the third quarter of 2025 down to 25% by the end of 2026.