The Defense Production Act and the Pentagon’s Anthropic pressure campaign, explained

Defense Secretary Pete Hegseth has reportedly given Anthropic a blunt choice: open its Claude AI technology for unrestricted military use by Friday or risk losing its government contract and being labeled a supply chain risk, reports the Federal News Network. Pentagon officials have also floated an escalation that would turn a contract dispute into a test of government power: invoking the Defense Production Act (DPA) to compel cooperation.

What the DPA is: Enacted in 1950 during the Korean War, the DPA is the president’s primary statutory tool to expedite and expand access to industrial resources deemed necessary for “national defense,” a term Congress has broadened over time to include certain emergency preparedness and critical infrastructure activities. The law is not a single “takeover” authority. It is a set of levers that, depending on the title used, can shift private-sector production and contracting priorities toward government needs.

How it works in practice: The best known authority is Title I (Priorities and Allocations), which allows delegated agencies to require US companies to prioritize certain government orders and allocate materials, services, and facilities to support national defense. Another major lever, Title III (Expansion of Productive Capacity and Supply), supports investments such as purchases, purchase commitments, and loans to expand or sustain production capacity.

Why the Anthropic angle is controversial: The Pentagon’s theory is that the DPA could be used to force changes that remove contractual “guardrails” or compel tailored delivery of AI capabilities—even where the vendor objects on safety or ethics grounds. Experts warn that using the DPA to dictate product terms this way would be novel and likely litigated.