OpenAI staff back Anthropic in Pentagon dispute

OpenAI staff back Anthropic in Pentagon dispute

By Gayane Tadevosyan
·2 min read

Anthropic warned in a court filing that its escalating dispute with the United States Department of Defense could cost the company up to $5 billion in lost business. The startup is asking a court to allow it to continue working with Pentagon-related partners while the legal battle unfolds.


More than 30 researchers from OpenAI and Google submitted a joint amicus brief supporting Anthropic. Among them was Jeff Dean. The researchers said they signed in a personal capacity and do not represent their employers.


In the filing, they warned that labeling Anthropic a “supply-chain risk” could harm the broader U.S. AI sector and weaken the country’s technological competitiveness.


The conflict began after negotiations between Anthropic and the Pentagon broke down over safeguards on how its AI models could be used, particularly concerning domestic surveillance and autonomous weapons. Last month, Defense Secretary Pete Hegseth said companies doing business with the U.S. military could not engage in commercial activity with Anthropic, significantly expanding the supply-chain risk designation.


Anthropic has since filed lawsuits in two courts, arguing the move violates its First Amendment rights and unfairly targets the company.


Executives say the fallout is already affecting finances. Chief financial officer Krishna Rao said hundreds of millions of dollars in expected Pentagon-related revenue are at risk this year, and losses could eventually reach $5 billion if partners are discouraged from working with the firm.


Chief commercial officer Paul Smith added that some partners have paused negotiations, demanded exit clauses, or canceled meetings due to fears about associating with the company.


The situation has drawn criticism across the tech industry. Sam Altman said enforcing the designation would be harmful for both the AI sector and the United States.


Despite the dispute, cloud providers such as Amazon and Microsoft say they will continue offering Anthropic’s Claude AI models to customers not connected to Pentagon contracts.


Anthropic is now seeking a temporary court order that would allow it to keep working with defense contractors while the case proceeds, with the first hearing potentially scheduled in San Francisco later this week. The Pentagon has not yet responded publicly to the latest filings.