Court strips Pentagon "risk" label from Anthropic, clearing a path back to defence work

US District Judge Rita Lin struck down the Pentagon's "supply chain risk" designation of Anthropic on 27 August 2026, removing a label that barred military contractors, suppliers and partners from doing business with the company. The ruling clears the way for Anthropic to re-enter defence-adjacent work that the designation had effectively frozen out.

Lin, of the Northern District of California, found the label unlawful on multiple grounds: retaliation against Anthropic in violation of the First Amendment, "arbitrary and capricious" decision-making, and a denial of due process under the Fifth Amendment. She said evidence of any genuine national-security risk was "slim", and pointed out the Pentagon kept trying to work with Anthropic even after slapping on the designation — undercutting the government's own case.

The dispute traces to Anthropic's refusal to let the Pentagon deploy its models for fully autonomous weapons or mass surveillance of Americans. Defense Secretary Pete Hegseth responded by declaring that no military contractor, supplier or partner could do business with the company. Until now, the "supply chain risk" tag had been reserved almost exclusively for firms tied to foreign adversaries; this marked its first use against a US company over a policy disagreement rather than a security concern, according to CNBC.

Anthropic filed suit in March 2026 and had already secured a preliminary injunction that month, meaning the label's practical bite was limited even before Thursday's final ruling. "We welcome the court's ruling that this supply chain risk designation was unlawful," an Anthropic spokesperson said. Lin was sharper still: "The empty invocation of national security is not a blank check to punish and retaliate against government critics," she wrote, according to NPR.

The government retains the option to appeal, and a second, related Anthropic lawsuit is still pending in Washington, DC.

The ruling is a win not just for Anthropic but for any AI vendor weighing whether to hold a safety line against a major government customer. A federal judge has now put on record that withholding capability on ethical grounds cannot be punished with an effective blacklist dressed up as a security finding — a precedent other labs will cite if they face similar pressure.

For the Pentagon, the loss narrows its leverage over frontier AI suppliers just as military demand for these systems grows. Watch whether Hegseth's department appeals, and how the pending DC case shapes the rules for future government-AI contracting disputes.