US appeals court sides 2-1 with the Pentagon on labeling Anthropic a supply-chain risk
On September 25 the DC Circuit denied Anthropic’s petitions 2-1, ruling the Defense Department did not overstep when it designated the company a supply-chain risk for restricting some military uses of Claude.
The US Court of Appeals for the DC Circuit ruled 2-1 on September 25 that the Defense Department could designate Anthropic a supply-chain risk. The dispute began when Anthropic refused to lift Claude’s restrictions on lethal autonomous weapons and mass surveillance.
The facts
- The reasoning: per Ars Technica, the majority held that the supply-chain-risk definition in 41 U.S.C. § 4713 covers “any person” and includes the act to “deny” use, with no bad motive required. Anthropic’s ability to disable Claude features qualifies.
- Two courts, two answers: last month a federal district court in Northern California found the designation illegal under a different statute, 10 U.S.C. § 3252, which requires malicious sabotage by an adversary. The appeals court didn’t dispute that finding. It said it was reviewing the broader authority in § 4713, over which it has exclusive jurisdiction.
- The dissent: Judge Karen Henderson argued the statute targets deliberate sabotage or eavesdropping, not a contractor openly enforcing usage restrictions.
- What’s next: Anthropic said it disagrees and is considering all options, including further review. It can seek en banc rehearing or petition the Supreme Court.
Our take
For companies and developers using Claude, the near-term effect is small: the designation covers federal agencies and defense contractors. The precedent is bigger. A vendor’s built-in usage restrictions can themselves be treated as a supply-chain risk.
If you pick models for government or defense work, availability now depends on more than capability and price. It also depends on how the vendor’s usage terms fare against the government’s demands. The case isn’t final, so watch whether Anthropic seeks rehearing.