Anthropic has lost a legal battle to overturn one of the supply chain risk designations imposed on the company by the Pentagon. On Friday, a federal appeals court in Washington, D.C., rejected the company’s challenge, allowing the designation to remain in place.

“The Department has sufficiently supported its conclusion that Claude’s continued integration into the Department’s information systems by the Department or its contractors poses a legally covered national security risk,” the justices wrote in the majority opinion. “As Anthropic has acknowledged, the company has encoded restrictions into Claude that prevent the model from performing tasks that Anthropic wants to prevent.”

Anthropic spokesperson Daniel Cohen said the company remains confident in its position and is considering all options. Those options could include an appeal to a broader panel of the D.C. Circuit Court of Appeals or to the U.S. Supreme Court.

Why the Pentagon designated Anthropic a supply chain risk

Earlier this year, the Pentagon sanctioned Anthropic under two separate supply chain laws. The designations would remove the company’s Claude AI model from the military and other federal government branches by this month.

Anthropic executives have said the company will not allow governments to deploy its current AI models to support autonomous weapons or domestic surveillance. Secretary of Defense Pete Hegseth deemed that position a serious national security risk.

Anthropic’s designations had to be challenged in separate courts. A federal judge in San Francisco revoked one of the supply chain risk designations in March and affirmed that decision last month. Friday’s ruling means the other designation remains in place indefinitely, allowing the Pentagon’s block on Anthropic to continue.

Anthropic says the designation affected business

Immediately after the designation, Anthropic said it lost revenue because customers were concerned about doing business with a government pariah. The company has not provided further updates on the designation’s effect on revenue.

Anthropic has continued to tout overall sales growth in recent months and is moving toward a potential initial public offering later this year.

Pentagon considers alternatives to Claude

The Pentagon has not provided detailed updates on its progress in replacing Claude with alternative AI models, including SpaceX’s Grok, Google’s Gemini and OpenAI’s GPT model.

Some employees at Google and OpenAI have objected to their employers’ agreements with the U.S. military. Anthropic rejected those agreements, citing ethical concerns. Both companies nevertheless proceeded with their arrangements and said support from the U.S. government was important.

Appeals court rejects Anthropic’s legal arguments

Friday’s ruling by the U.S. Court of Appeals in Washington, D.C., was somewhat expected. In April, the same three-judge panel declined to temporarily block Anthropic’s supply chain risk designation after finding that the company had not met the “stringent requirements” for an immediate moratorium.

During a hearing before the ruling, the three judges appeared to disagree over whether to block the designation altogether, challenging arguments from both Anthropic and the U.S. government. The final ruling was 2–1.

The majority also rejected Anthropic’s claims that the government violated its due process and free speech rights. The judges said the government had followed proper procedures and characterized the dispute as a standard contract negotiation.

The Pentagon “excluded Anthropic from its supply chain not based on the company’s support for increased government regulation of AI technology, but on the company’s refusal to agree to contract terms deemed essential by the Pentagon,” the judges wrote.

Anthropic had argued that the government acted beyond the authority granted under the Supply Chain Act. For now, however, the Pentagon and other Trump administration officials can continue to exclude Claude as Anthropic prepares for its expected IPO.