Amodei Slams OpenAI's Pentagon Deal as 'Safety Theater'
Leaked memo reveals Anthropic CEO's sharp criticism of OpenAI's DOD partnership and claims Pentagon disfavors his company.
Anthropic CEO Dario Amodei didn't hold back in a leaked internal memo on Friday, torching the deal between OpenAI and the Department of Defense as pure "safety theater."
The memo, obtained by The Information, shows Amodei telling employees that the Pentagon's chilly relationship with Anthropic stems partly from the company's refusal to offer what he described as "dictator-style praise to Trump."
It's a remarkably blunt assessment from the head of one of the biggest AI labs in the world — directly calling out rival Sam Altman's DOD arrangement while framing Anthropic's Pentagon struggles as a consequence of political dynamics rather than technical shortcomings.
The comments highlight the increasingly tangled relationship between leading AI companies and government contracts, where safety credentials and political alignment appear to be competing currencies.