WASHINGTON-A federal appeals court upheld the Defense Department's decision to designate Anthropic a security risk, giving the Trump administration a win in its complicated feud with one of the nation's top artificial-intelligence developers.

The federal appeals court in Washington said in a 2-1 decision Friday that the law giving the Pentagon the authority to deem a company a supply-chain risk gave Defense Secretary Pete Hegseth the latitude to take the action against Anthropic because the company's refusal to have its AI tools used in mass surveillance or autonomous weapons could undermine military operations.

"The Department had ample support for its conclusion that the continued integration of Claude into the Department's information systems, by the Department or its contractors, presented a statutorily covered national-security risk," the court wrote.

Anthropic had argued that the designation was an unprecedented attack retaliating against the company for disagreeing with the Defense Department in contract negotiations. A federal judge in California recently sided with the company in a parallel dispute covering parts of the designation under a different statute, fueling uncertainty about the future of the supply-chain risk label. That ruling made clear that other parts of the government couldn't blacklist Anthropic while the Pentagon fight continued.

Legal analysts have said that a split between the courts could ultimately lead to the Supreme Court.

"We respectfully disagree with the court's decision. Another federal court has already held the government's parallel designation unlawful. We remain confident in our position and are considering all options, including further review," an Anthropic spokeswoman said.

The supply-chain risk is normally used on companies based in foreign adversaries like China's Huawei that pose security risks. AI analysts have said the administration's move was a self-inflicted wound in the global AI race because it could damage Anthropic, which makes some of the nation's most powerful models and has said it has lost contracts and suffered reputational harm from the move.

A win for the government could give the Pentagon more latitude to take similar actions against companies in the future, analysts have said.

"By green-lighting the Pentagon's circumvention of standard procurement procedures to target Anthropic in this fashion, this ruling should alarm any government contractor," said Matt Schruers, CEO of the Computer & Communications Industry Association.

The rulings are part of a delicate relationship between the Trump administration and Anthropic, which are working together on the release of new AI models while the legal fight continues. The company has vaulted ahead in the AI race by many measures and is barreling toward an initial public offering in which it is aiming for a valuation of about $2 trillion.

Pentagon officials have said the department is nearly entirely done moving away from Anthropic's models, using tools from OpenAI and other competitors.

At a hearing at the federal appeals court earlier this year, the three-judge panel appeared split, with two of the judges seeming sympathetic to the administration's argument that it had the authority to use the supply-chain risk designation regardless of the guardrails Anthropic was pursuing in their negotiations.

"It's about risk, and they say, 'Well, based on what we know, we can't trust that the [AI] model may not have something embedded within it that is going to create a problem for military capabilities," said Neomi Rao, a judge appointed during the first Trump administration.

Judge Karen LeCraft Henderson, who was appointed by former President George H.W. Bush, disagreed and published a dissenting opinion. "To me this is just a spectacular overreach by the department," she said during the hearing.