The Pentagon is moving on from Anthropic. Following the collapse of a high-profile $200 million contract, the Department of Defense has begun engineering its own large language models, effectively ending its relationship with the AI company.

According to Pentagon Chief Digital and AI Officer Cameron Stanley, the work is already underway. "The Department is actively pursuing multiple LLMs into the appropriate government-owned environments," Stanley told Bloomberg. "Engineering work has begun on these LLMs, and we expect to have them available for operational use very soon."

The partnership fractured over fundamental disagreements on use. Anthropic insisted on contractual guardrails to prevent its AI from being used for mass domestic surveillance or in autonomous weapon systems. The Pentagon refused those limits.

With that impasse, the Defense Department turned elsewhere. It has since inked agreements with OpenAI and, for classified systems, Elon Musk's xAI for its Grok model. The shift aligns with a harder line from Defense Secretary Pete Hegseth, who has formally labeled Anthropic a supply-chain risk—a designation typically applied to foreign adversaries that prohibits Pentagon contractors from working with the firm. Anthropic is now contesting that label in court.

The message is clear: rather than reconcile, the Pentagon is building a future for its AI operations that does not include Anthropic's technology.

Source: TechCrunch