Background
Anthropic develops Claude, a general-purpose artificial intelligence model, and had supplied it to U.S. intelligence and defense agencies since 2024. The Department of War had used Claude Gov through partner platforms since March 2025. Federal defense contractors also provided products and services that integrated Claude or were created using Anthropic’s technology. The court’s account of those facts appears in the Northern District of California’s August 27, 2026 summary-judgment order.
The dispute centered on two restrictions in Anthropic’s Department of War usage policy: no mass surveillance of Americans and no lethal autonomous warfare. During negotiations over expanded military use, Anthropic agreed to remove most restrictions but declined to remove those two. The company treated them as contractual safeguards, while the Department of War sought an “all lawful uses” policy. The government’s stated rationale and characterization of Anthropic’s restrictions appear in its March 17 filing.
Under 10 U.S.C. § 3252, a supply-chain-risk designation addresses risks such as sabotage, maliciously introduced functions, or other conduct affecting covered national-security systems. The statute also requires findings that the designation is necessary to protect national security, that less intrusive measures are not reasonably available, and that specified procedural safeguards have been followed.