Pentagon Labels Anthropic an 'Unacceptable Risk to National Security'

The U.S. Department of Defense has filed a rebuttal to Anthropic's lawsuit, formally labeling the AI company an "unacceptable risk to national security."

In a new court filing, the DoD argues that Anthropic could "attempt to disable its technology or preemptively alter the behavior of its model either before or during ongoing warfighting operations" if the company felt its ethical red lines were being crossed — which the Pentagon deems an unacceptable risk.

The dispute began when Defense Secretary Pete Hegseth ordered all federal agencies to "immediately cease" using Anthropic's Claude AI in late February. Agencies including the Department of Health and Human Services and the General Services Administration removed Claude within hours. Anthropic responded with a lawsuit, arguing the "supply chain risk" label — previously only applied to foreign companies — was misapplied.

Nearly 150 retired federal and state judges have now filed an amicus brief supporting Anthropic, arguing the Pentagon "misinterpreted the statute and violated the necessary procedures."

Despite the legal battle, Anthropic continues to post record revenue figures. Industry monitors Epoch and Semianalysis estimate the company reached $2.5 billion in monthly revenue by February 2026 — putting it on track to surpass OpenAI by year's end. TIME Magazine has called Anthropic "the most disruptive company in the world."

The case is now before a federal judge. The ruling could set a major precedent for how the U.S. government can bar domestic AI companies from public contracts based on ethical principles.