Cult shortstop Burry: "Anthropic has left Palantir out of business in the AI market"

US financier Michael Burry used fresh data on corporate spending on artificial intelligence to bolster his long-standing bet against software developer Palantir Technologies. The investor, whose stance against the real estate market formed the basis of the movie "The Down Game," called Anthropic the clear leader in the battle for the corporate segment of the AI market.

"Anthropic is eating Palantir's lunch (from eat one's lunch - to beat the competition)," Benzinga quoted Burry's April 8 post in X, which has since been deleted. He pointed to Anthropic' s "tremendous jump" in annual subscription revenue (ARR) and attributed it to the company offering businesses a "simpler, cheaper and more intuitive solution."

"Palantir can settle for the public sector, which has low margins and small volumes. Anthropic's revenue has grown from $9 billion to $30 billion in a matter of months, while it took Palantir 20 years to reach $5 billion. Anthropic takes 73% of all new corporate spending, according to Ramp," Burry said.

Ramp's corporate expense management service recorded a record level of AI adoption in business in February 2026 - 47.6% of enterprises used it. At the same time, the share of Ramp customers with Anthropic subscriptions increased from 4% to 25% in just one year.

Anthropic now beats OpenAI, which created the ChatGPT chatbot, by about 70% in direct bidding for primary enterprise contracts - a "complete reversal" from 2025, when OpenAI itself was building market share faster than competitors, Ramp noted.

This reinforces Burry's thesis that Anthropic is now taking the bulk of new AI budgets, Benzinga believes.

At the end of trading in New York on Wednesday, April 8, when Burry's post was published, Palantir's shares collapsed by 6.2% to $140.8. The pressure factor was not so much Burry's words as the announcement of a new cloud service Managed Agents from Anthropic, which allows creating independent "AI-employees", StockStory points out. At trading on April 9, Palantir's securities were down another more than 7%.

Commenting on the Substack platform on the fall in Palantir's stock price, Burry refuted the suggestion that his tweet triggered the sell-off. "I don't believe my tweet brought down the stock. The reason was the potential end of the war and then the news about Anthropic. I just happened to write a tweet in the midst of all this," the financier said. "I was as shocked as anyone when I later saw the price drop to $140," he added.

During this discussion, Burry praised Anthropic's business model, calling the company "smart and well-managed." According to him, the startup could be generating profits right now if it decided to slow down the rate of investment in research and development (R&D). Unlike competitors that offer the market exclusively premium large language models (LLMs) - "a Ferrari-class product" - Anthropic gives a choice of "both Ferrari, Cadillac and Buick", allowing customers to optimize their cost structure, the financier said.

Burry also noted that Anthropic leases rather than builds its own data centers and adapts its products to run on Google's Tensor processors and Amazon's Trainium chips, reducing its reliance on Nvidia solutions.

In November 2025, Burry disclosed a large bet on the decline of Palantir's stock. He bought put options maturing in 2027, giving him the right to sell 5 million of the company's securities at $50 apiece. The investor calls Palantir a low-margin consulting business masquerading as a fast-growing AI company. Among the red flags he highlights are Palantir's problems in getting real money from clients and significant employee equity payments leading to share dilution, AInvest pointed out.

Burry also accused Palantir of "over-promising" in the artificial intelligence market. According to him, the company does not have its own AI engine and relies on external solutions like Anthropic's products. The financier claims that Palantir's government contracts have become less lucrative and its business is increasingly dependent on neural network developers, who now win corporate tenders on their own.

This article was AI-translated and verified by a human editor