Michael Burry, the investor famous for shorting the US housing market in 2008, spoke again over the weekend about his bets in the AI sector. The message: he’s keeping his puts against Palantir — and he’s calling Anthropic the far more interesting play.
The math he’s running
Burry’s argument is simple. Anthropic jumped from $9 billion to $30 billion in valuation and run-rate revenue within a few months. Palantir took two decades to reach $5 billion. His conclusion: Anthropic is the “easier, cheaper, intuitive solution for businesses.”
Palantir, by contrast, depends on government contracts — and those are “low margin and small.” Even with Trump’s recent supportive comments about Palantir, Burry is holding his puts.
What’s new here
Burry has warned about an AI bubble repeatedly over the past months. Him explicitly praising Anthropic is a new distinction: he’s separating overvalued incumbents from genuine challengers. His line on Anthropic: new companies are increasingly preferring Anthropic over established competitors — and that’s moving the market.
That fits the Ramp index I wrote about recently: the gap to OpenAI among paying US businesses is down to 4.6 percentage points. Burry is reading the same data as enterprise buyers and reaching the same conclusion.
My take
Burry is a contrarian, and you shouldn’t take his calls at face value. His Palantir short has been running for months without much to show for it. But the Anthropic praise is notable because he’s publicly committing to a thesis no one took seriously a year ago: that Anthropic can actually catch OpenAI in enterprise.
When even short-sellers switch the narrative, that’s a quiet but clear market signal. For Claude users: the product you use daily has reached the Wall Street phase where it’s no longer a forecast — it’s a valuation.
Sources: