Michael Burry said markets should crash hard enough to stop OpenAI and Anthropic from going public, calling the damage a public good. The investor made the remarks on X, where he agreed with a follower's joke that tanking the market would keep 'Skynet' from listing — a nod to the Terminator films' rogue AI. His posts landed the same week he swapped short positions for put options on several AI-linked stocks.
Burry's argument: the labs are burning capital
Burry's case rests on the numbers coming out of the labs themselves. Anthropic's prospectus shows a $42 billion net loss for 2025, alongside $518 billion in planned cloud and infrastructure obligations. BeInCrypto's breakdown put the operating loss closer to $8.06 billion. Read together, those figures describe a company planning to spend hundreds of billions while running deep in the red, and Burry's reply to his own thread was blunt: the companies will absorb and then destroy trillions of dollars in capital, and he called that the least of the damage.
He's been building this position for weeks. On September 14 he posted that the labs' pleas for a slowdown were self-serving. By Monday he'd moved from shorts to puts on several AI-linked stocks — a bearish bet with a defined expiry, not just a public complaint.
Anthropic's listing math
Anthropic could seek a valuation above $2 trillion, according to its prospectus as seen by Reuters. That's the number Burry appears to be arguing against: a debut priced at that level would require public markets to swallow a company with a 2025 net loss in the tens of billions and cloud commitments larger than most sovereign budgets. Reuters previously reported that Anthropic's debut is likely to slip past the November US midterm elections.
OpenAI's timing problem
OpenAI is on a different clock. It filed confidentially in June and now aims for a 2027 listing. CEO Sam Altman has called this year a poor time to list — a rare note of caution from the company's own leadership that lines up awkwardly with Burry's thesis, even if the reasoning differs. Burry says a crash would be good for the public. Altman says the window is bad. Both point the same direction: no IPO soon.
Even Grok agrees
When asked, Grok — xAI's chatbot — described a dot-com-style bubble built on unprofitable large language models falling short of AGI. Burry publicly agreed with that assessment. It is not every day that a sitting investor endorses a rival company's chatbot as a market analyst, but the exchange fits the pattern: Burry has spent the week collecting arguments, from his own posts to a follower's Terminator joke to a competitor's AI, and pointing them all at the same conclusion.
What happens next is mostly a matter of calendars. OpenAI's confidential filing keeps it on track for 2027. Anthropic's prospectus is out there with a $2 trillion figure attached and an election in between. Burry's puts have an expiry that hasn't been disclosed. Whether the crash arrives before the listings do remains the open question — and Burry is now betting on it.
