The ‘Big Short’ Guy Shuts Down Hedge Fund Amid AI Bubble Fears

The Big Short guy is short on trust in the state of the current economy. Earlier this week, it came to light that Michael Burry, famously depicted by Christian Bale in Adam McKay’s telling of the 2008 housing crisis, is shutting the doors to his hedge fund Scion Capital because he believes the market is completely detached from fundamentals.

In a letter dated October 27 informing his investors of his plans to liquidate the fund, he wrote, “My estimation of value in securities is not now, and has not been for some time, in sync with the markets.†The move comes shortly after Burry publicly shared that he had short positions against Trump administration-aligned defense tech firm and surveillance-state facilitator Palantir and chip giant Nvidia. On X, Burry revealed that his position has a $50 target for Palantir in 2027, despite the company currently trading at over $170 per share at the time of publication.

Burry came out of hibernation last month to speak up about what he believes to be a bubble in the AI sector, which he’s certainly not alone in ringing the alarm bells about. But Burry does have a novel angle on the matter: He believes companies are cooking the books by fudging the depreciation schedules of their chipsets, calling it “one of the more common frauds of the modern era.†Burry’s math is basically that hyperscalers like Oracle, Meta, Google, and others are claiming the useful lifespan of their tech is five to six years when, in his estimation, they are actually closer to two to three years.

That apparently pushed him to the conclusion that also ultimately led to him shutting down Scion Capital: “Sometimes, we see bubbles. Sometimes, there is something to do about it. Sometimes, the only winning move is not to play.â€

The depreciation tables are a bit of a weedsy observation, but one with significant impact if Burry is correct. He shared a chart that suggests earnings from these companies will be overstated by $176 billion between 2026 and 2028, with companies like Oracle and Meta overstating their actual earnings by more than 20%. He’s not alone in this observation, either. Another short seller, Jim Chanos, pointed to cloud-based GPU provider CoreWeave as a prime example of the deprecation trick, claiming that the company is barely profitable with more accurate deprecation timelines.

Any time a short seller shares information like this, they’re met with skepticism about their motivations. Palantir CEO Alex Karp took some shots at Burry’s position against his company, for instance, saying, “As far as I can tell, the two companies he’s shorting are the ones making all the money, which is super weird,†and arguing, “The idea that chips and ontology is what you want to short is batshit crazy.â€

But these guys aren’t the only ones starting to raise questions about how exactly these AI firms, seemingly endlessly taking on investor funds and racking up massive valuations, will actually make money. Source News reported that investors on a private call asked OpenAI’s Chief Financial Officer, Sarah Friar, why the company’s growth appears to be slowing, which she ultimately blamed in part on the dwindling time users were spending with ChatGPT following the addition of stronger content restrictions.

It’s not clear if “we’re building the erotica bot to maximize user engagement†is the kind of thing that should make the markets rest easier.