Pentagon officials and defense contractors are pushing back against an order to dismantle Anthropic’s AI systems, including its Claude model, following Defense Secretary Pete Hegseth’s March 3 directive labeling the company a supply-chain risk.
The ban mandates a six-month phase-out across all Defense Department networks and contractors. Yet many IT professionals inside the Pentagon say the move undermines productivity, calling it shortsighted and disruptive.
Anthropic secured a $200 million defense contract in July 2025 and became the first AI provider approved for classified military networks. Users report Claude outperforms alternatives like xAI’s Grok, particularly in consistency and coding tasks.
"Career IT people at DoD hate this move because they had finally gotten operators comfortable using AI," said an IT contractor. "They think it’s stupid."
Replacing Claude could take 12 to 18 months per system due to recertification requirements for secure environments, according to Joe Saunders, CEO of RunSafe Security. Some teams have reverted to manual data analysis using tools like Microsoft Excel.
Palantir’s Maven Smart Systems, used for intelligence and weapons targeting, relies heavily on Claude Code. Transitioning away will require rebuilding workflows and retraining AI agents, sources confirm.
Despite compliance orders, some agencies are slow-rolling the removal, betting the dispute will resolve before the deadline. The tension underscores a broader clash between rapid tech adoption and political oversight.