US Military Sticks With Claude as Other Defense Clients Bail

Anthropic finds itself in a peculiar and challenging position following its public disagreement with the Department of Defense. The company's artificial intelligence models are currently being deployed in the ongoing conflict between the United States and Iran, even as Anthropic simultaneously works to disengage from numerous clients within the defense sector.

This complex situation stems largely from a series of overlapping and often contradictory directives issued by the U.S. government. While President Trump had instructed civilian federal agencies to cease utilizing Anthropic's products, the company was granted a six-month window to conclude its engagements with the Department of Defense. However, the very next day, a joint surprise attack by the U.S. and Israel on Tehran initiated a sustained conflict, preempting the full implementation of Trump's directive.

Consequently, as the United States continues its aerial operations against Iran, Anthropic's sophisticated models are actively informing numerous targeting decisions. Despite Secretary of Defense Pete Hegseth's public commitment to classify the company as a supply-chain risk, no formal measures have been enacted to date. This absence of official action means there are currently no legal impediments to the continued operational use of Anthropic's systems.

Further insights into the deployment of Anthropic's technology emerged from a Washington Post article published on Wednesday. The report detailed how Anthropic's systems are being integrated with Palantir’s Maven system. As Pentagon officials formulated strike plans, these combined systems "suggested hundreds of targets, issued precise location coordinates, and prioritized those targets according to importance," according to the Post. The article characterized the system's primary function as providing "real-time targeting and target prioritization."

Concurrently, a significant number of companies within the defense industry have already begun to substitute Anthropic's models with alternative solutions from competitors. A Reuters report indicated that prominent defense contractors, including Lockheed Martin, initiated the replacement of the company’s AI models this week. Subcontractors are facing a similar dilemma; a managing partner at J2 Ventures informed CNBC that 10 of his portfolio companies "have backed off of their use of Claude for defense use cases and are in active processes to replace the service with another one."

The foremost unresolved question remains whether Secretary Hegseth will proceed with the supply-chain risk designation, an action that would almost certainly precipitate a contentious legal battle. In the interim, one of the leading artificial intelligence laboratories is progressively being phased out of military technology applications, even as its tools remain actively deployed in an ongoing war zone.