OpenAI Reaches Pentagon Agreement to Deploy AI Models on Classified Network

OpenAI said Friday it reached an agreement with the U.S. Department of Defense to allow its artificial intelligence models to operate within the department’s classified network under specified technical safeguards.

Chief Executive Sam Altman disclosed the arrangement in a post on X, stating the contract incorporates restrictions on domestic mass surveillance and requires human responsibility for the use of force, including in autonomous weapons systems. He said those principles are reflected in existing law and policy and have been incorporated into the agreement.

Altman said OpenAI will develop technical controls to govern how its models function within the defense environment. The company also plans to place engineers alongside Pentagon personnel to support model deployment and safety oversight. According to reporting by Fortune, Altman told employees at a company-wide meeting that OpenAI will be permitted to maintain its own “safety stack” to prevent misuse and that if a model declines to perform a task, the government will not compel the company to override that refusal.The agreement follows unsuccessful negotiations between the Defense Department and rival AI firm Anthropic. The Pentagon had sought commitments from AI providers permitting use of their models for “all lawful purposes,” while Anthropic sought to establish limits related to domestic mass surveillance and fully autonomous weapons.

Anthropic Chief Executive Dario Amodei said in a public statement that the company did not object to specific military operations but argued that certain AI applications could, in limited circumstances, undermine democratic values. The company said it had not received formal notice regarding the status of its negotiations with the Defense Department and stated it would challenge any designation as a supply-chain risk in court.

President Donald Trump criticized Anthropic in a social media post and directed federal agencies to phase out use of the company’s products over six months. Defense Secretary Pete Hegseth said Anthropic was seeking undue influence over U.S. military operational decisions and announced that companies doing business with the military may not engage in commercial activity with the firm, citing supply-chain risk concerns.

More than 60 OpenAI employees and roughly 300 employees at Google signed an open letter urging their employers to align with Anthropic’s position on safeguards. Altman said OpenAI has encouraged the Defense Department to offer similar contractual terms to other AI providers and expressed support for resolving disputes through negotiated agreements rather than litigation or government action.

The announcement was made shortly before reports that U.S. and Israeli forces had begun military operations targeting Iran, with Trump calling for the overthrow of the Iranian government.

About OpenAI

OpenAI was founded in 2015 and is headquartered in San Francisco, California. The company develops artificial intelligence systems, including large language models and generative AI tools used across commercial, enterprise, and government applications.

At the time of publishing, Stocks.News does not hold positions in companies mentioned in the article.