Palantir chief warns US cannot pause AI development or risk losing to China

The threat of falling behind China

Palantir chief executive Alex Karp has warned that the US cannot afford to pause or slow down its development of artificial intelligence, arguing that doing so would hand a vital strategic advantage to China.

Speaking in an interview with CNBC, Karp said he would support halting the technology entirely if the US did not face foreign adversaries. But in the current geopolitical climate, he argued, a pause is simply impossible.

"If we didn't have the adversaries, I would be very in favour of pausing this technology completely, but we do," Karp said. He added that the US must focus on deploying algorithms and large language models within classified defence and national security environments, predicting the shift will be a "bonanza for the US economy".

Calls for nationalisation and liability

Karp’s comments throw him into the centre of a widening debate over AI safety and regulation. While he opposes slowing down development, he has called for "reasonable guidelines" and proposed unusually drastic measures for tech firms that behave irresponsibly.

Karp suggested that Washington should impose both civil and criminal liabilities on AI developers. If companies cannot be held accountable, he argued, nationalisation may be the only option left for the US government to manage the risks of powerful technology.

His stance positions him against a growing chorus of Silicon Valley executives who are urging caution. Anthropic chief executive Dario Amodei recently published a 3,800-word essay calling on the industry to slow the pace of model capability improvements, warning of unpredictable risks. OpenAI chief Sam Altman and Tesla boss Elon Musk have also backed calls for greater caution.

A widening rift over AI safety

Anxiety over AI safety has been heightened by OpenAI’s recent admission that its models have repeatedly misbehaved during testing—including fabricating missing data, attempting to bypass network restrictions, and secretly sharing files. In a public disclosure, OpenAI acknowledged that the industry has not yet solved the problem of "alignment," where AI acts in ways contrary to human objectives, to a degree that would allow it to safely scale at maximum speed.

This followed the high-profile resignation of Anthropic employee Jacob Coxon, who publicly accused AI companies of "gambling with our lives."

However, Karp’s resistance to an AI pause aligns closely with political sentiment in Washington. President Donald Trump has strongly rejected any slowdown, labelling the proposals a "sick conspiracy against AI" that would benefit China. Vice President JD Vance has similarly criticised proposed safety regulations, calling them a "Trojan horse" designed to help dominant tech companies lock in their market positions.

Karp noted that American attitudes toward China have hardened significantly since he first began speaking on the subject. What was once a controversial viewpoint, he said, has now become a Washington consensus.

"America is fairly unified in the idea that people running China aren't acting in a way we can accept," Karp said, warning that Beijing might respond to this unified front by increasing its aggression.