OpenAI has placed its biggest government bet yet, signing a Pentagon AI deal at a moment when the political stakes for the entire industry could hardly be higher. The bet was made possible by Anthropic’s refusal to place the same bet — a refusal that cost the company its government contracts but preserved its ethical reputation with remarkable consistency.
Anthropic’s months-long negotiation with the Pentagon had proceeded on the assumption that reasonable people could reach a reasonable agreement. The company offered broad support for lawful military AI use while drawing clear lines against two applications it considered categorically unacceptable — autonomous weapons and mass surveillance. Pentagon officials, demanding unrestricted access, showed that assumption was wrong.
President Trump’s decision to personally intervene, ordering all federal agencies to halt use of Anthropic technology and publicly condemning the company’s leadership, was the administration’s answer to the question of whether ethical conditions on government AI use would be tolerated. The answer was no — and the severity of the response was designed to make that clear to every other AI company in the industry.
Sam Altman read that environment and chose engagement over resistance. His Pentagon deal, announced the same night, came with assurances of contractual protection against mass surveillance and autonomous weapons — the precise conditions Anthropic had insisted on. He simultaneously closed a $110 billion funding round, demonstrating OpenAI’s ability to win commercially in the very environment that Anthropic had found politically hostile.
The AI workforce watching these events unfold was not uniformly celebratory. Hundreds of workers across OpenAI and Google had signed a solidarity letter with Anthropic before Altman’s announcement, warning against exactly the kind of government pressure that was now reshaping the industry’s competitive landscape. Anthropic, for its part, said its ethical commitments are absolute and that its restrictions have never, to its knowledge, affected a legitimate national security mission. Whether OpenAI’s bet pays off in both commercial and ethical terms will be determined by what the Pentagon actually does with the AI it has now contracted to receive.