Anthropic Faces Customer Exodus and Billions in Losses After Government Supply-Chain Risk Ban

Anthropic signed a contract worth up to $200 million with the US Department of Defense in July 2025 to integrate Claude into military systems. The two sides later fell out over restrictions on the use of Claude for autonomous weapons and mass surveillance of Americans. The dispute raises questions about government procurement authority, corporate free-speech rights and the safety boundaries of military AI.

The Defense Department formally designated Anthropic a “supply chain risk” on March 5. On March 9, the company filed separate lawsuits in federal court in California and the US Court of Appeals in Washington, D.C. Judge Rita Lin suspended the designation and Claude ban on March 26. The D.C. court declined to stay the designation on April 8, and oral arguments concluded on May 19.

Founded by former OpenAI researchers in 2021, Anthropic focuses on its Claude models and enterprise AI. Claude Code has brought generative AI into software development and become a major revenue driver. The company’s refusal to allow its AI to be used for mass surveillance or fully autonomous weapons has put it at odds with the U.S. Department of Defense, with implications for government procurement and AI safety rules.

A March 3, 2026 report showed Anthropic’s annualized revenue had surpassed $19 billion, more than double the $9 billion recorded at the end of 2025 and up from about $14 billion just weeks earlier. The company completed a $30 billion Series G funding round on February 12 at a post-money valuation of $380 billion. The Pentagon formally designated Anthropic a supply-chain risk on March 5, but Claude added more than 1 million users per day that week.

Former OpenAI members Dario and Daniela Amodei founded Anthropic in 2021 with a focus on AI safety. The company secured a U.S. Department of Defense contract worth up to $200 million in July 2025. Its restrictions barring Claude’s use for mass domestic surveillance and fully autonomous weapons without human control have now put its technology governance policies at the center of defense procurement and industry competition.

While developing Claude 3.7 Sonnet, Anthropic delayed the model after testing indicated it could improve biological-weapons planning capabilities. The model was released on February 24, 2025. Its Cybench problem-solving rate within five attempts was about one-third, compared with only about 5% a year earlier. On February 27, 2026, Trump ordered federal agencies to stop using it, and the Defense Department designated it a national security supply-chain risk on March 4.