The relationship between Silicon Valley and the Trump administration just took a dramatic turn, one with potentially epic consequences for national security, the AI race, and the independence of American business leaders.
To summarize the fast-moving news story of the week: The Pentagon had a contract with Anthropic for $200 million. The deal let the military use its AI model Claude under two conditions: Claude couldn’t be used to power fully autonomous weapons—the kind that could decide to make a lethal strike without a human in the loop—and it couldn’t be used for mass surveillance of American citizens.
When Anthropic was founded, it was intended to build a “safer AI” than OpenAI, where Amodei had previously worked: He and his cofounders wrote a safety “constitution” for Claude, and these two measures, they felt, would breach it.
Secretary of War Pete Hegseth and President Trump were livid at the prospect of their power being limited by Anthropic. (It doesn’t help that some of Trump’s tech policy advisors, including “AI czar” David Sacks, have long been agitated by Anthropic, accusing it of being run by “woke” AI alarmists.) The feud spilled into public view when Undersecretary of War Emil Michael, a former sharp-elbowed Uber executive, put Amodei on blast on Thursday, accusing him of having a “god complex,” and saying in other remarks, “You can’t have an AI company sell AI to the Department of War and [not] let it do Department of War things.”
On Friday afternoon, President Trump escalated the feud further with an all-caps-studded message on Truth Social directing every federal agency to drop use of Anthropic. That’s a potential crippling blow to Anthropic’s business prospects, given that federal contracting is a nearly $800-billion-a-year market.
Hegseth went further: He designated Anthropic’s models “a supply chain risk,” and claimed that this would mean all companies doing business with the U.S. government would need to stop using Anthropic’s tech, even in cases when they were working on non-government projects.
The response is unprecedented against an American company, equivalent to a political assassination attempt against Anthropic. If fully executed, Trump and Hegseth’s demands would essentially demolish Anthropic’s enterprise business, which has been growing rapidly. At a minimum, they’ll sizeably dent Anthropic’s prospects for an IPO in 2027.
Anthropic has said it will sue the government to try to overturn Hegseth’s decision. In the meantime many companies may decide it is too much of a legal risk to keep using Anthropic’s AI.
In a stark reminder of just how cut-throat the battle for AI dominance has become: Fortune’s Sharon Goldman reports that, barely an hour after Trump’s announcement, OpenAI CEO Sam Altman told staff at an all-hands that their company was negotiating its own deal with the Department of War, stepping into Anthropic’s place. The deal successfully closed last night—with OpenAI agreeing to the Pentagon’s contractual terms, but also seeming to keep in place essentially the same two safety sticking points Anthropic fought to have.
The bigger questions here, though, are about democratic accountability and the relationship between government and business in America.
Who should decide how the U.S. government uses AI: The Pentagon—which, as defense-tech entrepreneur Palmer Luckey and others argue, represents a democratically elected government, and faces adversaries in China or Russia who won’t play by any AI rules in times of war? Or the AI maker, which understands the technology’s power and limitations best? “Congress is not the fastest-moving body in the world. And for right now, we are the ones who see this technology on the front line,” Amodei told CBS News.
The other big issue is whether it is OK for a democratic government to attempt to effectively destroy a business just because that company declines to agree to contract terms, or because it simply doesn’t like the politics of its CEO.
Wherever you fall on the spectrum, it’s not a situation to take lightly. And the major assault launched by the U.S. and Israel against Iran early this morning underscores just how high the stakes are.
Follow Alyson on X, LinkedIn, TikTok, Instagram, and the Titans and Disruptors vodcast.
