Pentagon Blocks Anthropic From Defense Contracts Over AI Safety Concerns

The Pentagon classified Anthropic a supply chain risk in May 2026, barring federal contractors from using Claude in military work. The decision reflects a fundamental tension: Anthropic's stricter safety controls versus the Pentagon's need for operational flexibility. Seven competitors—OpenAI, Google, Microsoft among them—gained access to classified networks, signaling that military requirements now trump the cautious approach some AI makers champion.