OpenAI awarded $200 million US defense contract

The declaration is estimated to implicit successful July 2026.

OpenAI is officially connected the Pentagon’s payroll. The Department of Defence announced a $200 cardinal declaration with OpenAI to supply the US authorities with caller artificial quality tools, including those utilized for proactive cyber defense.

In a station outlining its latest contracts, the DoD said that OpenAI “will make prototype frontier AI capabilities to code captious nationalist information challenges successful some warfighting and endeavor domains.” The enactment volition chiefly beryllium completed successful the Washington, DC portion with an estimated completion day of July 2026.

OpenAI said successful a caller blog post that the DoD declaration is its archetypal concern nether a caller inaugural to supply its AI exertion to workers crossed federal, state, and section governments. The institution is offering customized models for nationalist information connected “a constricted basis,” according to the announcement, with OpenAI saying that each usage cases indispensable comply with its policies and guidelines. OpenAI’s existent usage policy bans its services from being utilized to “develop oregon usage weapons” and “injure others oregon destruct property.”

“This contract, with a $200 cardinal ceiling, volition bring OpenAI’s industry-leading expertise to assistance the Defense Department place and prototype however frontier AI tin alteration its administrative operations, from improving however work members and their families get wellness care, to streamlining however they look astatine programme and acquisition data, to supporting proactive cyber defense,” OpenAI said.

This isn’t the archetypal clip OpenAI has bedded down with the military, having entered a concern with Anduril Industries successful December 2024 to integrate its AI bundle into the defence tech company’s counterdrone systems. The caller one-year DoD declaration is antithetical to earlier versions of OpenAI’s presumption of work that banned its exertion from being utilized for “military and warfare” — a prohibition removed by the institution past year. 

OpenAI’s DoD declaration is the latest of respective efforts to bolster subject efforts with AI technology, with the US having renamed the AI Safety Institute earlier this period to absorption connected combating nationalist information risks alternatively of wide safety. Rival AI developer Anthropic announced an AI exemplary with looser guardrails for US defence and quality agencies connected June 5th, Google removed commitments successful February to not usage AI successful ways “that origin oregon are apt to origin wide harm,” and Meta started allowing the US authorities to usage its Llama AI exemplary for “national information applications” past year.