Defense Secretary Pete Hegseth has given Anthropic CEO Dario Amodei until Friday evening (Feb. 27) to grant the military unfettered access to its AI model, Claude, or face serious consequences.
According to Axios, the 45-year-old former Fox News host warned during a tense meeting on Tuesday (Feb. 24) that the Pentagon could either cut ties and label Anthropic a “supply chain risk” or invoke the Defense Production Act to compel the company to tailor its model to military needs without safeguards.
He told Amodei he would not allow any company to dictate the terms under which the Pentagon makes operational decisions or object to specific use cases.
“The only reason we're still talking to these people is we need them and we need them now. The problem for these guys is they are that good,” a Defense official said ahead of the meeting.
A senior Defense official described the session as "not warm and fuzzy at all," while another source said it remained "cordial" with no voices raised and that Hegseth praised Claude directly to Amodei.
Anthropic has said it is willing to adjust its usage policies for the Pentagon but refuses to permit mass surveillance of Americans or the development of weapons that operate without human involvement.
During the meeting, Hegseth raised the Pentagon's claim that Anthropic had voiced concerns to Palantir about the use of Claude during the Maduro raid. Amodei denied that the company raised such concerns or discussed the matter beyond routine operating conversations, and reiterated that Anthropic's red lines have never hindered Pentagon operations or personnel in the field.
“During the conversation, Dario expressed appreciation for the Department's work and thanked the Secretary for his service,” a spokesperson told the outlet. “We continued good-faith conversations about our usage policy to ensure Anthropic can continue to support the government's national security mission in line with what our models can reliably and responsibly do.”
As explained by Axios, the Defense Production Act allows the president to require private companies to prioritize contracts deemed necessary for national defense. It was previously used during the COVID-19 pandemic to boost vaccine and ventilator production.
Officials are considering using it in an adversarial manner to force Anthropic to remove safeguards, though the company could challenge such a move in court.
Claude is currently the only AI model operating inside classified military systems. It was also used in the operation to capture Venezuela president Nicolás Maduro in January through Anthropic's partnership with Palantir.
Elon Musk's xAI has signed a contract to bring its Grok model into classified settings, but it is unclear whether it could fully replace Claude. The Pentagon has reportedly also accelerated talks with OpenAI and Google about moving their models into classified systems.
A source told Axios that Claude remains ahead in several military relevant applications, including offensive cyber capabilities, while Gemini is viewed as a possible replacement if Google agrees to allow use of its model for "all lawful purposes," which are the same terms Anthropic has rejected.