What Changed on September 25
The US Court of Appeals for the DC Circuit denied Anthropic's petition for review on September 25 and upheld the Department of War's designation of the company as a supply chain risk, in a 2-1 ruling written by Judge Gregory Katsas and joined by Judge Neomi Rao.
The decision closes a case that started March 9, when Anthropic first challenged the designation, and that moved through an expedited stay denial in April and oral argument on May 19. It is the second major court ruling this year on the same underlying contract fight, and it lands the opposite way from the first.
Two Cases, Same Contract Dispute, Opposite Results
Both designations trace back to one disagreement: the Pentagon wanted Claude available for "all lawful uses," and Anthropic refused to drop the restrictions it builds into Claude against autonomous lethal weapons and domestic mass surveillance. From that single refusal, two separate government designations followed, reviewed in two separate courts, under two separate legal theories.
| Case | Court | Filed | Decided | Legal theory | Outcome |
|---|---|---|---|---|---|
| Military-contract designation | N.D. California (Judge Rita Lin) | 2026 | Aug 27, 2026 | First Amendment retaliation, Fifth Amendment due process | Vacated |
| Civilian-contract designation | D.C. Circuit Court of Appeals | Mar 9, 2026 | Sep 25, 2026 | First Amendment viewpoint discrimination, arbitrary-and-capricious review | Upheld, 2-1 |
The California ruling this desk covered on August 28 vacated one of those two designations outright, on the grounds that it was retaliatory rather than a genuine security finding. It did not touch the second designation, which is the one the DC Circuit just upheld.
Why the Same Facts Produced Opposite Rulings
Anthropic ran the same core argument in Washington that had worked in California: that the government punished it for a policy position, not for a real security defect. The DC Circuit rejected that framing directly. "The Department excluded Anthropic from its supply chain based on the company's refusal to assent to a contract term that the Department deemed essential, not based on the company's support for greater governmental regulation of AI technology," Katsas wrote, distinguishing a contractual refusal from protected speech.
On the underlying security question, the majority gave the government wide latitude. "We have no reason to doubt that Anthropic manipulates Claude's function, use and operation with noble intentions," Katsas wrote, "but at least as applied here, the statutory definition of a 'supply chain risk' turns on what Anthropic does, not why Anthropic does it." Judge Karen LeCraft Henderson dissented, warning that this reading pressures every contractor to accept government demands or risk being labelled a security threat.
The Lesson for Anyone Fighting a Government Designation
A designation built on the same facts can survive review under one legal theory and fail under another, in different courts, at the same time. Treat each parallel proceeding as its own bet with its own odds, not as one fight that gets settled by whichever ruling lands first.
National-security deference is the government's strongest ground, stronger than the procedural and motive arguments that succeeded in California. A contractor weighing whether to push back on a government customer's terms should expect that even a documented pattern of retaliation in one venue will not automatically reach a differently-framed designation in another.
Read next: A King Asked AI's Leaders for Reassurance, Not Rules | No AI Safety Rule Binds Your Vendor Yet
