Anthropic's court win over Pentagon: Judge Lin rejects supply chain risk, backs Claude

At a July 30 hearing in San Francisco, Anthropic PBC faced off against the U.S. Department of War over the government's supply chain risk designation that would bar federal agencies from using its Claude AI models. Judge Rita F. Lin, who had already issued a preliminary injunction in March blocking President Trump's stop-use order, showed increasing skepticism toward the government's position. Department of Justice attorney James Harlow defended the designation, arguing that the government's actions concerned its own IT systems. But Judge Lin pressed on the lack of evidence that Anthropic could sabotage Claude after delivery, and called the Department of War's lost trust over Anthropic's refusal to adopt 'all lawful use' terms 'troubling.' Anthropic's legal team, led by Michael Mongan of WilmerHale, argued for summary judgment.

The hearing centered on whether the Pickering framework—typically balancing government employees' speech rights against employer interests—applies when the government punishes a contractor beyond termination. Harlow conceded that if the government made an example of a contractor to deter others, Pickering would apply. Judge Lin also questioned the secondary boycott scenario, hinting that the government's actions might exceed its contracting authority. Meanwhile, a separate case pending in the D.C. Circuit offers Anthropic another path, though that court denied injunctive relief. Congress is also stepping in: next year's defense appropriation bill adds language prohibiting the designation of domestic companies as supply chain risks for declining contract terms. With no factual disputes for a jury, Judge Lin is expected to rule on the record, potentially reshaping how the federal government can regulate AI suppliers.