Pentagon Resistance Mounts Against Hegseth’s Ban on Anthropic AI

The U.S. Department of Defense is entering a period of significant internal friction following Defense Secretary Pete Hegseth’s March 3 decision to designate Anthropic as a supply-chain risk. This move, which effectively bars the use of the company’s Claude AI models across the Pentagon and its sprawling contractor network, has triggered a rare level of open dissent among career IT professionals and military operators. The core of the dispute lies in a fundamental disagreement over AI guardrails; Anthropic’s insistence on specific safety constraints reportedly clashed with the Pentagon’s operational requirements, leading to a breakdown in the relationship and the subsequent blacklisting of the firm.

This regulatory pivot is not merely a change in vendor but a disruption of a deeply integrated technological ecosystem. Anthropic’s footprint within the defense establishment is substantial, anchored by a $200 million contract signed in July 2025. More importantly, Claude holds the distinction of being the first large language model (LLM) approved to operate on classified military networks. For operators who have spent the last year integrating these tools into workflows for targeting, operational planning, and intelligence analysis, the sudden mandate to transition represents a significant step backward in capability. IT contractors have been vocal in their criticism, noting that while the administration may prefer alternatives like xAI’s Grok, the technical reality is that Claude currently offers a level of consistency and reliability that its competitors have yet to match.

“

Anthropic’s footprint within the defense establishment is substantial, anchored by a $200 million contract signed in July 2025.

The logistical hurdles of this transition are immense. Replacing a core AI model in a military environment is not as simple as switching software licenses. Any replacement system must undergo a rigorous recertification process to ensure it meets the stringent security standards of the Department of Defense. Experts suggest that this process could take months, if not longer, creating a capability gap during the transition. The six-month phase-out period mandated by Hegseth is viewed by many within the Pentagon as overly optimistic, leading to a culture of 'foot-dragging' where staffers are delaying the removal of Anthropic tools in hopes that the policy will be reversed or modified before the deadline.