Palantir CEO Alex Karp is warning that OpenAI and Anthropic face so much lawsuit risk that the government will have to take them over — a claim that is bad news for their investors if true, but that looks mostly like a sales pitch for Palantir.

• Karp's argument: companies feeding their data into Claude and ChatGPT are quietly handing over their intellectual property, and when they realize the value they lost, they will sue.
• He says that liability is unlimited, so the only fix is for the government to nationalize OpenAI and Anthropic, which would wipe out investors and could drag the whole market down.
• Salesforce's Marc Benioff and Anthropic's Dario Amodei openly bragged that Anthropic staff now skip Salesforce and just ask Claude for their sales data, which is exactly the kind of "value leak" Karp is describing.
• The doom talk lands a month before Anthropic's expected IPO and doubles as a pitch for Palantir's "keep your data in-house" approach, so the collapse warning is overblown fear-selling rather than a real forecast.
• Palantir itself looks solid: revenue is growing far faster than costs, the balance sheet is strong, and even at a very high price-to-earnings ratio the stock looks reasonably valued given its growth, with room to run toward the mid-$200s.

Outlook: Expect more of this AI-liability and data-ownership fight as the Anthropic IPO nears, with Palantir stock likely to keep grinding higher on its growth while a real nationalization of OpenAI or Anthropic stays highly unlikely.