Anthropic PBC warned a federal judge on Tuesday that it faces the potential loss of billions of dollars in revenue this year due to the Trump administration’s decision to label the artificial intelligence startup a national security risk.
The designation, which follows a breakdown in negotiations over AI safety guardrails, has sent shockwaves through the tech industry’s relationship with the Pentagon.
At a hearing in San Francisco, Anthropic attorney Michael Mongan described a rapid erosion of the company’s enterprise business.
Since the "supply-chain risk" label was applied last week, more than 100 enterprise customers have contacted the firm to express doubt about future collaborations.
Mongan cited specific losses, including a $50 million contract paused by a financial services firm and a 50% reduction in a contract with a fintech company.
The dispute stems from Anthropic’s refusal to waive safety "red lines" regarding the use of its Claude model for mass domestic surveillance or autonomous lethal weapons.
The Department of Defense (recently rebranded the Department of War) countered that the military must have the authority to use technology for "all lawful purposes."
The legal battle has drawn in other industry heavyweights.
Microsoft, a major investor in the sector, filed a brief supporting Anthropic’s request for a temporary block on the government’s moves.
Microsoft argued that the broad nature of the designation threatens to paralyze ongoing Defense Department contracting for a wide array of IT products and services.
U.S. District Judge Rita F. Lin responded to the startup’s plea for urgency by moving the scheduled hearing from April 3 to March 24, 2026.
While Anthropic sought a commitment that the government would avoid retaliatory executive orders before that date, Justice Department lawyers declined to offer any such assurance.