A high-stakes game of chicken between Anthropic and the Department of Defense reached a boiling point late Thursday when the artificial intelligence (AI) company rejected the Pentagon’s “last and final offer” regarding military use of its AI model, Claude.

By late Friday, President Donald Trump dropped a bombshell: He ordered all federal agencies “immediately” terminate their use of Anthropic technology.

“We don’t need it, we don’t want it, and will not do business with them again!” Trump said in a post on his Truth Social platform. “There will be a Six Month phase out period for Agencies like the Department of War who are using Anthropic’s products, at various levels.”

At the heart of the impasse is a $200 million Pentagon contract and a fundamental disagreement over where lawful use ends and ethical catastrophe begin. Anthropic CEO Dario Amodei said Thursday that the company “cannot in good conscience” comply with the Pentagon’s demand to remove restrictions on how its technology is deployed, specifically citing fears of mass surveillance and the development of fully autonomous weapons.

The rejection follows a tense Tuesday meeting between Amodei and Defense Secretary Pete Hegseth. According to sources familiar with the matter, Hegseth issued a blunt ultimatum – allow Claude to be used for “all lawful purposes” or face immediate contract cancellation.

The Pentagon’s threats go beyond mere finances. Hegseth warned that Anthropic could be designated a “supply chain risk,” a blacklisting typically reserved for companies linked to foreign adversaries. The Department is also reportedly considering invoking the Defense Production Act to force compliance, a move experts say would trigger a landmark legal battle over corporate versus state control of dual-use technology.

Adding to the intrigue, OpenAI CEO Sam Altman told staff in an internal meeting Thursday that the company is working on a deal that could help resolve the escalating impasse between Anthropic and the Pentagon, according to a Wall Street Journal report, citing people familiar with the matter.

While the Pentagon claims its latest offer included compromise language, Anthropic dismissed the gesture as hollow. In a public statement, the company argued the proposal was “paired with legalese that would allow those safeguards to be disregarded at will.”

“I believe deeply in the existential importance of using AI to defend the United States,” Amodei wrote in a blog post while maintaining autonomous killing machines and domestic spying remain “outside the bounds of what today’s technology can safely and reliably do.”

The Pentagon, meanwhile, has signaled it will not be dictated to by Silicon Valley. Chief spokesman Sean Parnell said that while the agency has “no interest” in illegal surveillance, it will not allow a private vendor to “jeopardize critical military operations.”

“Without a deal, Anthropic would be forced into offboarding from defense work,” said one source close to the negotiations, who spoke on condition of anonymity.

The standoff highlights a growing rift in the AI industry. While rivals like OpenAI, Google, and Elon Musk’s xAI have reportedly agreed to the “all lawful purposes” clause for their military contracts, Anthropic remains the only provider currently integrated into the military’s classified networks.

Public opinion appears to lean toward the startup’s cautious approach. A new survey by the Information Technology and Innovation Foundation (ITIF) found that 67% of Americans believe tech companies have a responsibility to set limits on their products, even if the government disagrees.

The Pentagon’s deadline of 5:01 p.m. Friday looms. If no deal is reached, the DoD has already begun vetting alternatives, reportedly asking major defense contractors like Boeing and Lockheed Martin to review their dependencies on Anthropic.

Amodei says the company is prepared to “offboard” and assist in a transition to other providers to avoid disrupting missions, but the fallout could set a massive precedent for the future of the American military-industrial-AI complex.