Dario Amodei’s Anthropic Sues US Government, Saying Pentagon Retaliated Over AI Safety Guardrails

Anthropic has taken the extraordinary step of suing the U.S. government after the Pentagon labeled the artificial intelligence (AI) firm a “supply chain risk to national security,” igniting a legal clash that could reshape how Washington controls artificial intelligence vendors working with the military.

Company executives argued that those guardrails were necessary because current frontier AI models remain too unreliable for autonomous weapons and because large-scale surveillance programs could conflict with constitutional protections. Anthropic stated the restrictions had never interfered with any military mission during its prior work with the government.

Soon afterward, Hegseth moved to classify Anthropic as a supply chain risk under federal procurement authority outlined in 10 U.S.C. § 3252. The designation, confirmed in a letter sent to the company around March 4, restricts certain government contractors from using technology considered vulnerable to sabotage or foreign influence.

Anthropic argues the move stretches the statute far beyond its intended purpose. In court filings, the company contends the designation was meant for foreign-linked security threats rather than policy disagreements with domestic suppliers. The lawsuits claim the government failed to use the “least restrictive means” required by law and instead imposed what the company describes as an unofficial blacklist.

The legal filings also raise constitutional concerns. Anthropic argues the designation violates the First Amendment by punishing the company for publicly advocating limits on AI uses such as autonomous weapons and domestic surveillance. According to the complaint, labeling the firm a national security risk damages its reputation and could threaten hundreds of millions of dollars in contracts.

Anthropic is asking courts to block enforcement of the designation, order federal agencies to withdraw directives halting business with the company, and prevent similar actions in the future. The company said its goal is not to force the government to purchase its technology but to prevent retaliation over policy differences.

Pentagon officials have not publicly commented on the litigation, citing policy regarding active court cases. Some defense leaders have previously argued that military agencies must retain full operational authority over contractor technologies during emergencies and cannot allow vendors to dictate how systems are used.

The dispute arrives as the race to secure military AI contracts intensifies. Rival firms, including OpenAI reached agreements with the Pentagon around the same time Anthropic negotiations broke down. Meanwhile, major technology partners such as Google and Microsoft have indicated they intend to continue working with Anthropic on commercial services unrelated to defense.

Industry analysts say the outcome of the case could establish a precedent for how the federal government pressures AI companies to modify safety policies when national security interests are involved. For now, Anthropic’s consumer products and commercial AI services remain available, while the legal battle over the Pentagon designation begins working its way through the courts.