Appeals court upholds Pentagon's blacklist of Anthropic, days after a different court called it unlawful

A federal appeals court ruled Friday that the Pentagon can continue blacklisting Anthropic as a "supply chain risk," upholding a designation that a separate federal judge had called illegal only a month earlier — leaving the company's legal fight split across two conflicting rulings.

The U.S. Court of Appeals for the D.C. Circuit ruled 2-1 that the Department of Defense 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." One judge dissented, arguing the government had not shown sufficient evidence that Anthropic's technology could be manipulated to create the risks the department cited.

Why the Pentagon blacklisted Anthropic

The dispute traces back to early 2026, when Defense Secretary Pete Hegseth and the Trump administration designated Anthropic a "supply chain risk" and ordered federal agencies — not just the Pentagon — to stop doing business with the company. The trigger was Anthropic's refusal to let the Pentagon use Claude for "all lawful purposes," a blanket authorization the company declined to grant because of its own policies against enabling autonomous weapons systems and surveillance of U.S. citizens.

A split verdict across two courts

Anthropic sued in March, arguing the designation was retaliatory rather than a genuine security assessment. It won a significant victory in August, when U.S. District Judge Rita Lin in Northern California ruled the label "arbitrary and capricious" and an unlawful First Amendment retaliation, finding also that the company had been denied due process under the Fifth Amendment. Judge Lin cited Pentagon statements criticizing Anthropic's "arrogance" for raising concerns about how its products would be used as evidence the designation was punitive.

Friday's D.C. Circuit ruling reaches the opposite conclusion in a related legal challenge, finding the designation was properly supported by the government's national-security rationale. "Another federal court has already held the government's parallel designation unlawful," an Anthropic spokesperson said in a statement. "We remain confident in our position and are considering all options, including further review."

What happens next

Anthropic can petition for an en banc rehearing before the D.C. Circuit's full 11-judge panel, or appeal directly to the Supreme Court. In the meantime, the blacklist upheld by Friday's ruling remains in effect within that court's jurisdiction, even as Judge Lin's contrary ruling stands in the separate California case — an unusual state of open legal conflict for a dispute this significant.

The case is shaping up as one of the first major tests of whether AI companies can maintain their own safety-driven limits on government customers — declining specific use cases like autonomous weapons or mass surveillance — without the government punishing them for the refusal. How it's ultimately resolved could influence whether other AI labs adopt similar restrictions on their own government contracts, or treat "sell to the government on the government's terms" as effectively non-negotiable.