Appeals Court Upholds Pentagon Blacklist of Anthropic

A federal appeals court has sided with the Pentagon in its fight with Anthropic. On 25 September a three-judge panel of the U.S. Court of Appeals for the D.C. Circuit voted 2–1 to uphold the Defense Department’s designation of Anthropic as a supply-chain risk, denying the company’s petitions to overturn it. The label bars the military, and contractors working on its systems, from using Claude.

The decision lands a month after the opposite result in California. In August, U.S. District Judge Rita Lin vacated one of the Pentagon’s two designations, finding it was First Amendment retaliation for Anthropic’s public criticism of the administration. The department had issued a second designation under the Federal Acquisition Supply Chain Security Act, 41 U.S.C. § 4713, and challenges under that law go straight to the D.C. Circuit. That is the one the appeals court has now left standing, so Anthropic won one case and lost the other.

Judge Gregory Katsas wrote the majority opinion, joined by Judge Neomi Rao. The core of it is the two restrictions Anthropic refused to drop from its contract: no use of Claude in fully autonomous weapons, and none for mass domestic surveillance. “On more than one occasion, these restrictions have stopped Claude from performing tasks requested by government users,” Katsas wrote, according to The Epoch Times. The department, he said, “had ample support for its conclusion” that keeping Claude in its information systems “presented a statutorily covered national-security risk”, and it “reasonably feared that Anthropic might manipulate Claude’s design” to stop it performing functions the department considered authorised. The panel rejected Anthropic’s arguments that the designation was arbitrary, exceeded the statute or punished protected speech.

The majority was explicit that it was not deciding who is right about AI weapons. “In our Republic, it is the President and the Secretary of War who must determine how best to balance the competing risks,” Katsas wrote. Judge Karen LeCraft Henderson disagreed on what the statute is for. The 2018 law, she argued, was written to keep “hostile nation states and other bad actors” out of federal systems, and in her dissent she wrote that “a contractor’s honest and upfront enforcement of restrictions” is not the kind of risk it lets the government act on.

The dispute dates to September 2025, when talks over putting Claude on the department’s GenAI.mil platform broke down. Anthropic had signed a $200 million Pentagon contract that July; the department then wanted access for all lawful purposes, and Anthropic wanted written assurance on its two red lines. The designations followed in February. Defense Secretary Pete Hegseth greeted the ruling with a post: “The @DeptofWar does what is right for the Country and our Warriors.”

Anthropic’s response pointed to the other case. “We respectfully disagree with the court’s decision. Another federal court has already held the government’s parallel designation unlawful,” a spokesperson told CNBC, adding that the company is “considering all options, including further review”. That could mean asking the full D.C. Circuit to rehear the case, or going to the Supreme Court. Either way, the practical position has reversed since August: the retaliation finding in San Francisco still stands, but the designation that actually keeps Claude out of defence systems is back on firmer ground, and it now has an appellate court behind it.