Anthropic Clashes with Pentagon Over AI Safeguards
Anthropic is reportedly in a standoff with the Pentagon over its refusal to relax safety guardrails for military applications of its AI models. The dispute threatens a $200 million defense contract, as Defense Secretary Pete Hegseth has allegedly demanded the removal of constraints to enable broader military deployment. This conflict occurs as Anthropic's Claude product line reaches a reported $14 billion revenue run rate and the company has raised $30 billion at a $380 billion valuation.
- The Pentagon's demand for "any lawful use" of AI stems from its January 2026 "AI Acceleration Strategy," which established seven "Pace-Setting Projects," including initiatives for autonomous swarms and AI-driven battle management, from campaign planning to kill chain execution. - During a meeting on February 24, 2026, Defense Secretary Pete Hegseth gave Anthropic CEO Dario Amodei a deadline of February 27 to remove its usage restrictions. The Pentagon has threatened to invoke the Korean War-era Defense Production Act (DPA) to compel the company to comply with its terms. - Anthropic's specific contractual guardrails prohibit the use of its models for autonomous weapons systems that lack human control and for mass surveillance of American citizens. - The Pentagon has initiated a "supply chain risk" assessment, asking major defense contractors like Boeing and Lockheed Martin about their dependence on Anthropic's technology. A negative designation could effectively force contractors like Palantir, Anduril, and Amazon Web Services to remove Claude from their offerings for the defense sector. - On the same day as the Pentagon meeting, Anthropic updated its "Responsible Scaling Policy," removing a key pledge to halt the training of more powerful AI models if their safety could not be guaranteed. The company cited the need to remain competitive as a reason for the change. - While Claude was previously the only model approved for use on the military's classified systems, the Pentagon recently signed a deal allowing Elon Musk's xAI to be used in classified settings without similar restrictions. The Department of Defense also holds contracts of up to $200 million each with Google and OpenAI. - The confrontation is viewed as a key test for the future of AI procurement and the relationship between national security agencies and Silicon Valley. The "all lawful purposes" standard is expected to become the default for all future defense AI contracts, creating a potential two-tier market for AI firms.
