Michael Burry, the investor who bet against the housing market before the 2008 crash, has a new target.
A research firm has pegged Anthropic’s value at $150 billion, about 92% below what the AI company is seeking in its initial public offering, and has called the listing the most absurd IPO of 2026, CNBC reported on Oct. 7. The same day, Burry joined the skeptics, calling it another bubble.
The disagreement is not about whether Anthropic has built something real. Its Claude models compete at the top of the market with OpenAI’s GPT. The dispute is about what that is worth, and whether the numbers in the prospectus support the price the company wants.
The prospectus shows a company still losing money, with computing commitments that stretch years into the future. The bull case rests on growth: if revenue keeps compounding, today’s losses become tomorrow’s margins. The bear case is that the growth is assumed rather than demonstrated.
The gap between the two views is extraordinary even by the standards of a heated market. A valuation of $150 billion is not a small number. That it could be described as 92% below the asking price says more about the asking price than about the research firm’s restraint.
Anthropic has ridden the same wave that lifted every leading AI lab. Founded by former OpenAI researchers, it built a reputation for safety and for Claude, a model family that has matched the frontier on many benchmarks. Private investors valued the company at successively higher marks, and the jump to a public listing was meant to be the moment the private market’s optimism was tested against the public market’s discipline.
The company’s pitch to public investors is that its models are gaining share and that the compute it has locked in will become an asset rather than a liability. That is a harder story to sell than the private market’s version, because public investors must reconcile the promise with a profit-and-loss statement that still shows red ink and a balance sheet committed to spending for years.
The broader IPO market has not been kind to AI companies that priced too high. Several listings have traded below their offer price, and the memory of those disappointments is fresh. A company that prices for perfection in that environment is asking investors to forget a lesson they just learned.
Burry’s skepticism carries weight in certain corners of the market. He is followed closely by investors who believe the AI rally has detached from the fundamentals, and his description of Anthropic as another bubble was quickly picked up by the same outlets that covered the research firm’s estimate. A single bearish call can move sentiment more than it should, and this one landed on the eve of a listing.
The timing is awkward. Anthropic has been racing to expand its product line and cut prices ahead of the listing, moves read by analysts as an effort to show momentum rather than just promise it. The harder the company pushes to demonstrate growth, the more scrutiny its numbers draw.
The core tension in AI valuations is that the technology is genuinely valuable and the prices are genuinely extraordinary. Both can be true at once. The question for Anthropic’s IPO is which one the public market will choose to believe, and whether private-market prices can survive contact with investors who mark things to market every day.
Companies that go public at a discount to their last private round face a re-rating, and the re-rating is rarely gentle. If public investors price Anthropic closer to the research firm’s figure than to the company’s target, the listing becomes a test of how far AI’s private valuations can travel before they meet the public market.
The skepticism extends beyond one company. If Anthropic, with a leading model and a deep bench of researchers, cannot justify its asking price in the public market, the implication is that no AI lab can. That is why the dispute over a single valuation has become a proxy for the larger question of whether the entire sector’s pricing is sustainable.
Burry has been early before and wrong before. His calls are followed more for the questions they raise than for their precision, and the question he is raising now is the one the whole market is quietly asking: how much of the AI boom’s value is real, and how much is a number someone wrote down.
Burry’s record is mixed, and a single research firm’s number is one data point. But the fact that the gap between the company’s asking price and an independent estimate can be measured in the hundreds of billions of dollars is its own signal. When a listing can be called the most absurd of the year and the accusation lands without being laughed off, the market is already asking a harder question than the company wants to answer.
