DC Circuit upholds Pentagon supply-chain-risk designation against Anthropic

A federal appeals court has handed Anthropic a defeat in its fight over a Pentagon blacklist. On September 25, the U.S. Court of Appeals for the D.C. Circuit ruled 2-1 that the Department of Defense was within its rights to designate Anthropic a supply-chain risk, rejecting the company's challenge and leaving Claude off-limits to the military and its contractors.

What's new

Bloomberg reported that the court "rejected Anthropic PBC's challenge to a Trump administration designation of the artificial intelligence company as a supply-chain risk." Judge Gregory Katsas wrote the majority opinion, joined by Judge Neomi Rao; Judge Karen LeCraft Henderson dissented. Katsas held that "the Department had ample support for its conclusion that the continued integration of Claude into the Department's information systems…presented a statutorily covered national-security risk," according to CNBC's report on the ruling.

Henderson's dissent argued the majority stretched the law past its intent, writing that it does not treat "a contractor's honest and upfront enforcement of restrictions" as the kind of supply-chain risk that justifies a blacklist.

An Anthropic spokesperson responded: "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." The panel delayed the ruling from taking immediate effect, giving Anthropic a window to seek a rehearing before the full D.C. Circuit or petition the Supreme Court.

Context

The dispute traces back to March 2026, when Defense Secretary Pete Hegseth designated Anthropic a national-security supply-chain risk. The rupture began the previous September, when talks over deploying Claude on the Pentagon's GenAI.mil platform broke down: the department wanted unrestricted access to Anthropic's models for any lawful purpose, while Anthropic wanted enforceable limits keeping Claude out of fully autonomous weapons and domestic mass-surveillance systems. Anthropic had already signed a $200 million Pentagon contract in July 2025 before that later negotiation collapsed.

Because the Pentagon relied on two separate designations against the company, the fight has played out in two courts at once. A federal judge in San Francisco struck down one of those designations in August. Friday's D.C. Circuit ruling concerns the second, and leaves it standing.

Why it matters

This is the most consequential legal setback yet for an AI lab that has staked its public identity on refusing certain military and surveillance use cases. Anthropic's argument, that the government retaliated against it for enforcing its own safety commitments, didn't move the majority, which sided with the Pentagon's national-security judgment over the company's contract terms. That leaves a frontier AI developer formally locked out of federal defense work by a nationally binding designation, with its argument having failed in the court that matters most for setting precedent.

The split outcome between the two courts, one striking down a nearly identical designation and the other upholding it, makes near-term resolution unlikely without further appeal. It also puts other AI labs on notice: a company's own safety guardrails, the restrictions it builds in specifically to avoid weapons or surveillance misuse, can themselves become the basis a government agency cites to justify shutting a vendor out entirely. Anthropic's stated next moves, an en banc rehearing request or a Supreme Court petition, mean the case is far from over, but for now, the Pentagon's blacklist against Claude stands.