President Donald Trump has ordered U.S. government agencies to stop using Anthropic's products, according to ABC News, while Defense Secretary Pete Hegseth designated the AI company a national security "supply-chain risk" — the most dramatic escalation yet in a monthslong feud between the Pentagon and the maker of Claude.

The moves followed a 5 p.m. Friday deadline in the company's standoff with the Department of Defense, and they land just days after Anthropic CEO Dario Amodei sat down with Trump at a private White House dinner on Sunday and appeared beside the president at Tuesday's signing of an industry-wide AI safety accord at the White House.

For continuing coverage of the unfolding dispute and every other major development, follow our breaking AI news coverage as the story develops.

Why the Pentagon Turned on Anthropic

At the center of the dispute is Anthropic's refusal to accept a Pentagon ultimatum to lift internal policies that restrict how its AI can be used by the military. According to Breaking Defense, the company declined to remove safeguards against the use of its AI for lethal autonomous weapons or mass domestic surveillance.

Anthropic representatives had opposed the use of the company's products for surveillance purposes or to develop lethal autonomous weapons, Reuters reported. That stance put the San Francisco-based lab on a collision course with a Pentagon that has spent 2026 rapidly onboarding commercial AI: the department's GenAI.mil platform initially contracted Google's Gemini, then OpenAI's ChatGPT, and Hegseth later added xAI's Grok for military use.

The conflict has already cost Anthropic business. The Department of Defense terminated a contract with the company estimated at roughly $200 million, according to reporting cited in a Wikipedia summary of the dispute, and rival labs absorbed much of the military work Anthropic declined to enable on the government's terms.

The Legal Battle So Far

The Pentagon first labeled Anthropic a supply-chain risk in March, a designation under the Federal Acquisition Supply Chain Security Act (FASCSA) that bars covered systems and contractors from doing business with the company. Anthropic sued the Trump administration shortly afterward.

The litigation has produced a split record:

  • March 2026 — Judge Rita Lin of the Northern District of California granted Anthropic a preliminary injunction, writing that the Pentagon designated the company a supply-chain risk because of its "hostile manner through the press" and calling the punishment "classic illegal First Amendment retaliation."
  • April 2026 — The D.C. Circuit denied Anthropic's emergency motion to stay the FASCSA designation, which remains in effect for covered systems while litigation continues.
  • August 2026 — Judge Lin ruled that the blacklisting itself was illegal.
  • September 2026 — A divided D.C. Circuit panel upheld the Pentagon's designation 2-1, refusing to block it, according to Reuters and CNBC.

The appeals court's decision was a significant legal defeat for Anthropic. The panel's order noted that lifting the designation mid-litigation "would force the United States military to prolong its dealings with an unwanted vendor of critical AI services in the middle of a significant ongoing military conflict."

A Feud With a Strange Choreography

The escalation sits awkwardly alongside this week's displays of detente. On Sunday, Amodei joined Trump for a private White House dinner — their first one-on-one meeting — after being conspicuously absent from the president's state dinner for Chinese President Xi Jinping the previous week. On Tuesday, Amodei stood flanked by Trump alongside Elon Musk, Mark Zuckerberg, Jensen Huang, Sundar Pichai and other tech executives as the industry signed a voluntary "super intelligence" safety accord that Trump described as "morally binding."

Yet the Pentagon's posture has only hardened. Axios reported that the Pentagon warned Anthropic it would "pay a price" as the feud escalated, and Hegseth's formal supply-chain-risk designation extends the practical effects far beyond the Defense Department: defense contractors that work with covered systems must now cut ties with the company, and Trump's new directive pushes civilian agencies in the same direction.

What It Means for the AI Industry

The dispute has become a test case for how far the federal government can go in punishing an AI company over its usage policies — and the answer, so far, is farther than most of the industry expected. Legal experts cited by Wired said Anthropic has a strong case, but noted that courts are often reluctant to overrule the White House on national-security matters.

The stakes extend beyond one company. If the government can designate an AI lab a supply-chain risk because of the safety limits it places on its own models, every commercial AI provider faces a choice between building customizable frontier systems for the military and accepting the kind of usage restrictions that Anthropic insists on. Anthropic's approach — publishing usage policies and enforcing them, even for powerful customers — has been widely cited as an industry benchmark.

Anthropic did not immediately detail its next legal steps after the D.C. Circuit ruling, and the White House has not publicly detailed the scope of the new directive beyond the order that agencies stop using Anthropic's products. A petition for rehearing by the full D.C. Circuit, or an appeal to the Supreme Court, would be the expected avenues for the company to challenge the designation.

For now, the message to Silicon Valley is unambiguous: in this administration's Pentagon, the price of saying no is losing the government as a customer entirely — and the biggest safety-focused lab in the industry has become the example.

Stay Ahead of AI

Washington's fight with Anthropic is reshaping how AI companies, the military and regulators interact. For more analysis and the latest AI developments from Washington, Silicon Valley and the research labs, make AI Buzz Wire your daily briefing.

Read more AI news →