Investors are increasingly eager to point out the flaws in the AI economy and to label Leopold Aschenbrenner's earlier investment frenzy as an act of madness. Yet they remain highly cautious and unwilling to move too soon to bet against it.
Steve Eisman, a former hedge fund manager and well-known short seller, has become one of the most prominent critics of the AI boom sweeping the market, which some argue even threatens the world. Eisman rose to fame for his trades during the 2008 financial crisis, and his story was chronicled in Michael Lewis's book The Big Short. But even though he is critical of the AI boom, he has been reluctant to put money behind a short bet. He told me, "I'm not ready to short heavily here yet." Right now he is focused intently on Nvidia, the core target of the AI trade, and on Nvidia's business relationships supplying chips to Anthropic and OpenAI. He said, "The entire supply chain is dependent on two companies that are deeply unprofitable." He is watching whether OpenAI can go public soon: "A company that keeps burning cash—companies bleeding this much money like these two—the narrative matters enormously because they have to keep raising capital." Eisman said that if OpenAI keeps delaying its IPO, that could become the trigger for him to short the AI sector.
Meanwhile, AI euphoria remains at a historic peak, with bubble talk constantly coming out of Silicon Valley and Wall Street. The discussion usually centers on absurdly high valuations, circular financing, off-balance-sheet loans, a lack of profits, and the enormous debt piling up around AI deals. A market boom built on such a fragile foundation tends to attract financial sharks looking to bet on what they see as an imminent crash. I had expected a large wave of short sellers eager to plow millions of dollars into shorting AI. But after talking to people, the real picture on Wall Street is far more complicated than imagined. A handful of short sellers have already positioned themselves, with Michael Burry, another Big Short prototype figure, leading the charge. Others are probing cautiously, trying to figure out how to carry out AI short trades. But AI bears have not charged in en masse. Part of the reason is that over the past two decades, a long bull market has left many short sellers badly beaten and deeply in the red. Trillions of dollars in capital have kept pushing stocks higher, and many believe it is too early to enter now and the cost is extremely high.
One hedge fund manager runs a thematic short position in the AI sector (a financial term meaning shorting an entire track), and he admits he acted too early. Since 2023, he has been cautiously shorting AI-related stocks. This year his trades started to turn around. The short seller, who asked not to be named, said, "We're doing well this year." He asked for anonymity because he worries about retaliation from the companies he is shorting.
After the blowup of Leopold Aschenbrenner's hedge fund Situational Awareness, the situation took a turn, which was a godsend for his short portfolio. He said, "Since then this short trade has performed very well." Aschenbrenner was heavily invested in a large number of AI stocks, with his top ten holdings including Nvidia, Oracle, SanDisk, Micron Technology, CoreWeave, and Nebius Group. The short seller commented, "A 24-year-old with no financial experience at all can still attract a flood of capital—that's probably the peak of madness."
One of the most discussed AI short targets in the market is the new cloud provider CoreWeave, which is almost the internet-famous concept stock of the AI track. CoreWeave's short interest has fallen from a 2026 peak of 26%, but short positions still account for as much as 15% of the float. Such a high short position has also led retail investors in the Reddit community to discuss the possibility of a short squeeze—once the stock price rises, short sellers will be forced to urgently buy shares to close out. This battleground stock has fallen by one-third over the past year, but in 2026, as short sellers clearly retreated, the share price rebounded somewhat. Matthew Unterman, head of research at short data firm S3 Partners, estimates that short sellers collectively lost about $460 million on CoreWeave this year. In September last year, short research firm Kerrisdale Capital published a short report on CoreWeave. At the time, the company had been public for only five months, and its stock had already risen 200%. The short firm described it as "a debt-driven GPU rental business with no moat, just packaged as an innovation story." But on the day the report was published, CoreWeave announced a partnership with Nvidia, and the stock surged in response. Many aggressive short sellers closed their positions and exited immediately after publishing their reports. Sam Adlagi, founder of Kerrisdale Capital, was unwilling to discuss the trade. He told me he was unsure of the final profit or loss on the short, and Kerrisdale no longer holds a short position in CoreWeave. (Now some short sellers have set their sights on its competitor Nebius, whose stock has soared this year.)
As for Oracle, it has been a major trophy for short sellers this year. George Noble, a former Fidelity fund manager who once profited from the bursting of the internet bubble, has publicly advocated shorting Oracle on X and in his Substack column; the stock has halved this year. Unterman estimates that short sellers collectively made about $2.3 billion on Oracle this year, and Oracle's short interest rose 70% during the year. However, short sellers account for only 2.6% of the float. Although Eisman himself has not yet shorted any AI stocks, he sold his Oracle holdings earlier this year. He believes Oracle is highly dependent on OpenAI and is in a precarious position. Combined with Oracle's heavy debt burden, short sellers point out that the company's credit rating was recently downgraded and is only one notch away from junk status.
Another approach to shorting focuses on the private credit market. Short seller Lakshmi Ganapathi, who just launched the bearish-strategy fund Minerva Investment Management, said she is not acting yet and is sorting through private credit exposure to the AI sector, trying to find the weakest link in the risk chain. She said, "A massive amount of private credit money is piling into this track." She noted that data center companies are borrowing heavily in the private credit market, and she is studying each company's loan structure one by one, the concentration of debt maturities, and whether loan quality is deteriorating.
Shorting AI ultimately comes back to Nvidia. After Nvidia's GPUs helped OpenAI launch ChatGPT in 2022, Nvidia's stock has risen more than 2,000%, and its revenue growth has been astonishing. Eisman closely tracks Nvidia's revenue to judge whether its growth has peaked. In the most recent quarter, Nvidia's revenue doubled year over year, higher than the previous quarter's 85% growth; however, 70% of revenue came from just five large cloud providers. He commented, "Quite astonishing—Nvidia's growth itself is the core argument of the AI bull narrative." But so far, shorting Nvidia has kept short sellers bleeding. S3 Partners data shows that Nvidia shorts are the biggest losers this year, collectively losing about $8 billion on the stock; it is the single stock with the largest short interest globally, with short exposure exceeding $62 billion. However, short sellers account for only 1.3% of the float, a ratio that has barely changed this year.
Although Burry has loudly proclaimed a short on Nvidia, not everyone has followed suit. Unterman said Nvidia's enormous short position includes hedging demand and is not entirely a pure one-sided bearish bet. In other words, some traders short Nvidia to hedge long positions in other tech stocks. Overall, short sellers seem to be waiting for the right moment to act. Short seller Mark Cohodes said, "Is AI a bubble? Without a doubt. Unless you're constrained, no one can fail to see this is a bubble, and the bubble will burst one day. But in the market, far too many people have fallen on the battlefield because they shorted too early."
Comments