Michael Burry, most famous for shorting the housing market ahead of the 2008 financial crisis, said that ‘Anthropic is eating Palantir’s lunch’ in a now-deleted post on X (1).
Since sharing that thought on April 8, Palantir has fallen rapidly — over 13% in the last five days. Although, the stock has recovered slightly after President Donald Trump made a Truth Social post in support of the company on April 10.
Must Read
Burry first announced he had put options on Palantir in the fall of 2025, which allow holders to sell a stock for a set price as long as it is sold by a certain date. A put is helpful if you think a stock is going to crash soon. In his newest Substack post, he says he’s maintained puts on Palantir ever since, and he’s not planning on selling them despite Trump’s recent posts (2).
Here are some reasons he might be shorting the company and why he thinks Anthropic is a better bet.
Burry says Anthropic’s target market is behind its fast-rising revenue
Elsewhere in his deleted post, Burry said, “Anthropic went from $9B to $30B in months, it took $PLTR 20 years to get to $5 Billion.” He says that’s because Anthropic’s product is the “easier, cheaper, intuitive solution for businesses.”
In comparison, he says that Palantir’s government contracts don’t set them up for the same quick growth. His post says that “PLTR can have government, which is low margin and small.”
He also says “Anthropic is taking 73% of all new enterprise spending per Ramp.” But that’s not quite accurate: The Ramp report says that 73.3% of new businesses choose Anthropic over OpenAI when they spend on AI for the first time (3). The cited numbers don’t include companies that chose other AI options, such as Google or xAI.
According to Ramp’s report, OpenAI is still the company with the most adopted model overall, with 34.4% of U.S. businesses having a paid subscription with the company. Anthropic is in second place at 24.4%.
Still, Anthropic’s comfortable second-place seat, along with new companies showing preference for it over the market leader, means that the company is in a good place to take over more of the AI market. This is especially true now that it’s announced Project Glasswing (4), which is an initiative to use the company’s new Claude model, Claude Mythos, to protect against AI-driven cyberattacks.