Joe Lonsdale co-founded Palantir, runs the venture firm 8VC and is an investor in Anthropic. On Friday, at Reuters’ Momentum AI event in Austin, he said out loud what plenty of people in the industry think and few will attach their name to: the leading AI companies’ warnings about existential risk are a dangerous attempt to shape public policy.
“I think that what OpenAI and Anthropic are pushing right now is very dangerous, and we don’t want an oligopoly that controls all of our policy here with the government,” he said. “Safety is a real thing, but fear and regulation, we’ve got to be really careful how we respond to those things.”
The accusation is more specific than the usual complaint
Lonsdale is not claiming the risks are invented. His accusation is a funding chain: people close to Anthropic and at OpenAI are bankrolling outside groups that advocate for stronger AI regulation. Write the rules and you entrench your position. That is the standard regulatory capture argument, and this time it lands on the companies positioning themselves as the cautious ones.
It fits the week we’ve had: a White House memo casting Amodei as the face of AI doomerism, and Europe’s tech bosses lining up behind the call to slow down. The fault line no longer runs between industry and government. It runs straight through the industry.
The investor praises what he attacks
The framing is what makes this interesting. In the same conversation Lonsdale goes out of his way to praise Amodei and Anthropic. The company has assembled exceptional talent and emerged as a leader in the race. “It’s very hard for anyone to second-guess the highest-performing company in the world over the last few years,” he said. “I’m proud to be an investor in Anthropic.”
He describes himself as a small investor there. He also said two things that would have been headlines in a quieter week: defence companies he backs will likely buy Ukrainian defence firms, and with autonomous weapons, accountability has to stay with military commanders. He did not put himself forward for the senior Pentagon role his name keeps getting attached to. His focus, he said, is his family and his businesses.
Both things can be true, and that’s the hard part
Amodei can sincerely believe the technology needs slowing down, and slowing it down can benefit Anthropic at the same time. Sincerity and self-interest are not mutually exclusive. That is exactly why Lonsdale’s point is harder to wave off than the phrase “very dangerous” suggests. It is also worth remembering that Palantir’s co-founder is not a disinterested bystander when the subject is regulating military AI. Nobody in this argument is without a book.