US Court Upholds Pentagon's 'National Security Risk' Label for Anthropic

A panel of judges from the District of Columbia Court of Appeals has ruled against AI developer Anthropic, refusing to pause a Pentagon designation that labels the company a national security supply chain risk. The court determined that the government's interest in securing AI technology during active military conflicts outweighs the potential financial or reputational damage to the firm. This designation is unprecedented for a US-based company. It effectively bars Pentagon contractors from utilizing Anthropic’s AI models, creating a significant barrier to federal adoption of the company's Claude LLM and restricting its footprint within the US defense infrastructure. The conflict originated from a July 2025 agreement intended to integrate Claude into classified networks. Negotiations collapsed in February 2026 when the US government demanded unrestricted military access to the technology. Anthropic resisted, maintaining that its AI should not be utilized for mass domestic surveillance or the development of lethal autonomous weapons. The legal battle intensified after President Donald Trump ordered federal agencies to cease using Anthropic products in late February. While a California district court previously issued a preliminary injunction against the directive, the DC Circuit's latest ruling maintains the statutory designation, which the court noted may cause 'irreparable harm' to the company but is justified by military readiness. This decision establishes a potentially chilling precedent for AI developers regarding government compliance. If the 'supply chain risk' label remains, it could signal a more aggressive regulatory stance toward AI firms that attempt to impose ethical constraints on military applications, potentially shifting the competitive landscape toward more compliant providers.