AI in court

Anthropic loses on appeal, and its usage limits now count as a military risk


The D.C. Circuit ruled 2–1 on September 25 that refusing military uses is enough to remove Claude from U.S. defense suppliers.

September 25, 2026 · Translated from the Spanish original

What happened

Why it matters

The number

180 days. The deadline the Department set in March to remove Anthropic’s products from its systems and bar defense contractors from using them.

Context

On August 27, a federal judge in California struck down a parallel designation issued under a different law and called it retaliation, as we reported when that trial-court ruling came out. The D.C. Circuit says it doesn’t disagree with that conclusion: the law applied in California requires ill intent by an adversary, and the 2018 law, according to the majority, doesn’t. Two courts, two laws, two opposite outcomes on the same facts.

What’s next

Bottom line

Anthropic came into this fight calling for slowing the pace of models and opening its systems to outside auditors. The ruling answers that the control it exercises over Claude is precisely what makes it a risky supplier.

Sources

Edited by Rodrigo Cornejo. How we select and verify: who writes these notes.

Related notes

← All notes