The U.S. Court of Appeals for the D.C. Circuit on April 8, 2026, rejected Anthropic's emergency motion to halt the Department of Defense's 'supply chain risk' designation against its Claude AI models. The ruling keeps the Pentagon's ban on federal contractors using Claude in force while litigation proceeds. The panel expedited oral arguments to May 19, 2026.
Background: The Clash Over AI Safety Guardrails
The designation stems from a breakdown in negotiations between Anthropic and Pentagon officials in late February 2026. The core dispute revolves around two restrictions in Anthropic's terms of service: a prohibition on fully autonomous weapons systems, including armed drone swarms operating without human oversight, and a ban on mass surveillance of U.S. citizens. Emil Michael, the Pentagon's chief technology officer, called these restrictions 'irrational obstacles' to military competitiveness, particularly against China. Programs like the Golden Dome missile defense initiative and the need for rapid hypersonic threat response were cited as critical use cases. Anthropic offered limited, case-by-case exceptions but refused to eliminate its core safety guardrails, citing reliability concerns for high-stakes autonomous decisions. After talks collapsed, President Trump directed all federal agencies to phase out Anthropic technology within six months, followed by Defense Secretary Pete Hegseth's supply chain risk designation—a measure typically reserved for foreign entities like Huawei.
Legal Duel: Two Courts, Conflicting Rulings
Anthropic filed parallel lawsuits in March 2026: one in the U.S. District Court for the Northern District of California, and another in the D.C. Circuit targeting the specific procurement statute governing supply chain risk. On March 26, U.S. District Judge Rita F. Lin granted Anthropic a preliminary injunction in the California case, ruling that the administration's actions appeared punitive rather than protective, lacked sufficient statutory justification, and overstepped authority. That order temporarily lifted the designation. The Trump administration appealed to the Ninth Circuit. The April 8 D.C. Circuit decision runs counter to Lin's ruling, creating a legal tension over enforceability. The D.C. Circuit panel—comprising judges Gregory Katsas and Neomi Rao, both Trump appointees—acknowledged Anthropic would 'likely suffer some degree of irreparable harm' but concluded the balance of equities favored the government, citing judicial management of AI technology security 'during an active military conflict.'
Industry Warnings and Wider Implications
Industry observers flagged the case as a warning for U.S. AI development. Matt Schruers, CEO of the Computer and Communications Industry Association, said the Pentagon's actions and the D.C. Circuit ruling 'create substantial business uncertainty at a time when U.S. companies are competing with global counterparts to lead in AI.' The case now moves toward the expedited May 19 oral argument in the D.C. Circuit, with the Ninth Circuit appeal still pending. The outcome will likely define the limits of federal power to designate domestic AI firms as national security risks and the extent the government can pressure private companies to alter their AI safety policies. Separately, Anthropic's unreleased Claude Mythos Preview has autonomously discovered thousands of high-severity zero-day vulnerabilities across all major operating systems and web platforms, including bugs in Linux and OpenBSD that humans missed for decades—a capability that underscores the tension between AI safety and military utility.