Anthropic says it can't agree to the military's AI use terms — then it got slammed by an official

• Anthropic is negotiating with the Defense Department over the terms of use for its frontier model.
• CEO Dario Amodei said in a statement that it will not budge on two red lines.
• Several hours after Amodei's statement, an official slammed him on X, calling the CEO a "liar."

Anthropic's CEO is prepared to walk away from its contract with the military, according to a new statement published on Thursday.

In a blog post, CEO Dario Amodei said that the company "cannot in good conscience accede" to the request of the Defense Department concerning safeguards around its frontier model, Claude.

On Tuesday, Defense Secretary Pete Hegseth gave Anthropic an ultimatum to agree with the military's terms over the use of Claude or get blacklisted by the government.

Defense officials gave Anthropic until Friday evening to agree to the terms.

The terms were not clarified, but the issue, according to Amodei's statement, appears to revolve around two red lines Anthropic is not willing to cross when it comes to how Claude is deployed: mass domestic surveillance and fully autonomous weapons.

A spokesperson for Anthropic declined to comment.

Hours before Amodei put out a statement, Sean Parnell, a Pentagon spokesperson, posted on X that the department had no interest in using AI to conduct mass surveillance of US citizens or to develop autonomous weapons.

And several hours after Amodei released his statement, the Department's undersecretary for research and engineering, Emil Michael, lashed out at the CEO in a social media post.

"It's a shame that @DarioAmodei is a liar and has a God-complex. He wants nothing more than to try to personally control the US Military and is ok putting our nation's safety at risk," Michael wrote on X.

"The @DeptofWar will ALWAYS adhere to the law but not bend to whims of any one for-profit tech company," added Michael, who was previously Uber's chief business officer.

A person familiar with the negotiations told Business Insider the department provided a new proposal just 36 hours before Hegseth's deadline, and the language around the provisions on mass surveillance and autonomous weapons allowed for "any lawful use" of Anthropic's AI.

The person said that the additions essentially gave the military to discretion to set aside Anthropic's red lines and use Claude as it sees fit.

A senior Pentagon official told Business Insider on Tuesday that the department will consider invoking the Defense Production Act — a wartime law that would essentially give the president control over Anthropic's resources in the interest of national security — and deem the company a supply chain risk.

Both uses of the national authorities would be unprecedented, experts told Business Insider, considering that the levers are being used as a negotiating tactic and against an American company.

"I'm not aware of this ever having been used as a weapon in a negotiating posture," Dean Ball, an ex-senior policy advisor for the White House Office of Science and Technology Policy, told Business Insider.

Amodei wrote in his blog post that the two threats are "inherently contradictory: one labels us a security risk; the other labels Claude as essential to national security."

The CEO wrote that Anthropic hopes the government will "reconsider" its position on the safeguards and that the company's preference is to continue working with the military.

"Should the Department choose to offboard Anthropic, we will work to enable a smooth transition to another provider, avoiding any disruption to ongoing military planning, operations, or other critical missions," Amodei wrote.

It's not yet clear how the Pentagon plans to respond.

February 27, 2026: This story was updated to reflect a public statement by Emil Michael, the Pentagon's undersecretary for research and engineering, about Amodei.