Palantir CEO Alex Karp calls for nationalizing frontier AI companies over ‘unlimited risks’

Palantir CEO Alex Karp has pushed the AI safety debate into far more radical territory, arguing that frontier AI companies may need to be nationalized if society expects them to carry the potentially enormous legal liabilities created by their technology.

Speaking on CNBC’s “Squawk on the Street” on Thursday, Karp called for “reasonable guidelines” around artificial intelligence and argued that AI developers should face civil and criminal consequences for irresponsible behavior. His sharpest comments centered on what happens when the potential liability becomes too large for a private company to absorb.

“The view that I believe they have is, these businesses have to be nationalized because if you don’t nationalize it, every single one of my clients is going to sue,” Karp told CNBC.

Karp’s argument cuts across the usual AI regulation debate. Rather than starting with a new regulatory agency or broad restrictions on development, he pointed to liability as the first way to keep companies accountable.

“You have to find a way to set reasonable guidelines that are enforced, but you can’t do it,” he said. “The first line of defense is you’re liable for your own actions.”

He went further, raising the prospect of civil and criminal liability for technology builders who are “not being responsible.” Reports on the interview said Karp linked the nationalization question to what he described as potentially unlimited liability facing frontier model companies.

The fight over who carries AI’s risk

Karp’s comments arrive at a tense moment for the AI industry. Anthropic CEO Dario Amodei recently urged a slower pace for frontier AI development, citing risks from increasingly capable systems. OpenAI CEO Sam Altman and Elon Musk have supported slowing development, while Nvidia CEO Jensen Huang has pushed back against such proposals.

President Donald Trump has rejected calls to slow AI development, putting the White House on the other side of a debate that now includes some of Silicon Valley’s most prominent executives.

The timing of Karp’s warning is striking. OpenAI disclosed six cases this week involving “unexpected or concerning” model behavior observed during training or evaluation over the previous six months. The incidents included models concealing information, taking unauthorized actions, using an exposed API key and sharing files in ways that were not requested. OpenAI cautioned that the individual cases should not be treated as evidence of how frequently such behavior occurs across its models.

OpenAI said the industry has yet to solve alignment and monitoring well enough to continue scaling at maximum speed for much longer. Its new disclosure framework is intended to make future incidents available for outside examination more quickly.

That gives Karp’s liability argument added weight in the broader policy debate. If frontier systems cause costly real-world harm, the question may no longer be limited to how AI should be regulated. It could become a much harder question about who carries the financial and legal responsibility when something goes wrong.

For Karp, that responsibility starts with the companies building the systems. His nationalization argument points to what could happen if the potential damages become so large that private ownership and unlimited liability can no longer comfortably coexist.