Ars Technica · AIJon BrodkinSep 26, 2026, 05:36 UTC+8

US appeals court rules Pentagon can blacklist Anthropic over refusal to open Claude features

Court rules Pentagon can blacklist Anthropic for refusing to enable Claude features

The US Court of Appeals for the DC Circuit ruled 2-1 that the Defense Department may blacklist Anthropic for refusing to open certain Claude features to the military, even without bad faith by Anthropic. The ruling said the case involves hard questions about military use of powerful AI, and found the Defense Secretary did not exceed his authority under the Supply Chain Security Act or the Constitution, so the petition for review was denied. The court had already rejected Anthropic's emergency stay request in April.

Why it matters: The ruling marks out how far the Defense Department can restrict AI suppliers under supply chain security law, in a fight over military use of AI.

Read the original ↗Export Markdown

Take

The court didn't rule on whether Anthropic was right or wrong. It ruled on whether the Defense Secretary overstepped — and said 2-1 that he didn't, so the blacklist stands. The line in the ruling that matters is the comparison: the military worries a tightly constrained model goes offline at the worst moment; Anthropic worries an unconstrained model points a lethal strike at the wrong target. Both risks are real. Handing the trade-off back to Hegseth means these fights get settled by the executive branch, not by judges.

For Anthropic, the loss isn't this case. It's the line it has tried to hold: model capabilities can be opened up by use case, tier by tier. That line doesn't hold in government procurement — either everything is open, or you're out. An emergency stay was already denied in April; this just makes the outcome final.

Source: Ars Technica · AI · arstechnica.com
Published: Sep 26, 2026, 05:36 UTC+8 · 4 days ago