OpenAI awarded $200 million US defense contract

OpenAI is officially on the Pentagon’s payroll. The Department of Defence announced a $200 million contract with OpenAI to provide the US government with new artificial intelligence tools, including those used for proactive cyber defense.

In a post outlining its latest contracts, the DoD said that OpenAI “will develop prototype frontier AI capabilities to address critical national security challenges in both warfighting and enterprise domains.” The work will primarily be completed in the Washington, DC region with an estimated completion date of July 2026.

OpenAI said in a new blog post that the DoD contract is its first partnership under a new initiative to provide its AI technology to workers across federal, state, and local governments. The company is offering custom models for national security on “a limited basis,” according to the announcement, with OpenAI saying that all use cases must comply with its policies and guidelines. OpenAI’s current usage policy bans its services from being used to “develop or use weapons” and “injure others or destroy property.”

“This contract, with a $200 million ceiling, will bring OpenAI’s industry-leading expertise to help the Defense Department identify and prototype how frontier AI can transform its administrative operations, from improving how service members and their families get health care, to streamlining how they look at program and acquisition data, to supporting proactive cyber defense,” OpenAI said.

This isn’t the first time OpenAI has bedded down with the military, having entered a partnership with Anduril Industries in December 2024 to integrate its AI software into the defense tech company’s counterdrone systems. The new one-year DoD contract is antithetical to earlier versions of OpenAI’s terms of service that banned its technology from being used for “military and warfare” — a prohibition removed by the company last year. 

OpenAI’s DoD contract is the latest of several efforts to bolster military efforts with AI technology, with the US having renamed the AI Safety Institute earlier this month to focus on combating national security risks instead of overall safety. Rival AI developer Anthropic announced an AI model with looser guardrails for US defense and intelligence agencies on June 5th, Google removed commitments in February to not use AI in ways “that cause or are likely to cause overall harm,” and Meta started allowing the US government to use its Llama AI model for “national security applications” last year.