Anthropic supply-chain designation upheld in AI military dispute

A federal appeals court on Friday upheld the Department of War's decision to exclude artificial intelligence developer Anthropic from its supply chain, marking a victory for the Trump administration in the ongoing dispute over military use of frontier AI models. The core issue revolves around Anthropic's refusal to grant the Department unrestricted access to its Claude model, despite relaxing public usage policies to accommodate classified tasks.

CEO Dario Amodei defended the restrictions, citing them as central to the company's identity. Tensions peaked in 2026 when Secretary of War Pete Hegseth demanded complete AI dominance, leading to a clash during a Venezuela operation over mission-critical system reliability. Anthropic argued its usage limits don't constitute supply chain risk under 2018 statutes, arguing they target foreign adversaries, not domestic partners.

The court rejected this, stating statutory definition broadly covers any disruption or denial of system functionality, regardless of motives. The panel also dismissed Anthropic's constitutional claims, ruling the post-deprivation process met Fifth Amendment standards given AI's rapid evolution. Circuit Judge Karen LeCraft Henderson dissented, warning the majority's broad interpretation transforms the statute from preventing foreign espionage into a tool to strip vendors of rights without evidence of subversive conduct.

The ruling's immediate impact on Anthropic's defense footprint remains unclear due to a parallel California lawsuit that temporarily lifts the ban. Despite legal uncertainty, Anthropic's commercial trajectory appears strong, with private sector demand driving enterprise valuation upward as Pentagon contracts shift to rival OpenAI.