OpenAI strikes landmark Pentagon deal to deploy AI models on classified network

Altman said OpenAI is urging the Defense Department to offer the same terms to all AI providers.

• OpenAI to deploy AI models on US military's classified network
• Deal bans mass surveillance and autonomous weapons
• OpenAI will provide engineers to ensure safety and compliance

Did our AI summary help?

OpenAI, on Saturday, announced reaching an agreement with the United States Department of Defense to deploy its artificial intelligence models within the military’s classified network. This marks a significant step in Silicon Valley’s deepening ties with the Pentagon.

Chief Executive Officer Sam Altman announced the deal in a post on microblogging site X early in the day, "Tonight, we reached an agreement with the Department of War to deploy our models in their classified network," Altman wrote, referring to the Defense Department. He said the Pentagon had shown "a deep respect for safety and a desire to partner to achieve the best possible outcome".

Altman outlined that two of OpenAI’s core safety principles -- prohibitions on domestic mass surveillance and maintaining human responsibility for the use of force, including in autonomous weapons systems -- were embedded in the agreement.

Tonight, we reached an agreement with the Department of War to deploy our models in their classified network. In all of our interactions, the DoW displayed a deep respect for safety and a desire to partner to achieve the best possible outcome. AI safety and wide distribution of…

— Sam Altman (@sama) February 28, 2026

"The DoW agrees with these principles, reflects them in law and policy, and we put them into our agreement," he said. "We also will build technical safeguards to ensure our models behave as they should, which the DoW also wanted," added Altman.

OpenAI said it would deploy field deployment engineers to support implementation and ensure safety compliance, and that its models would operate on cloud networks only.

Meanwhile, an OpenAI spokesperson, earlier this week, confirmed to CNN that the company maintains the same "red lines" as rival Anthropic when it comes to military work, specifically opposition to domestic mass surveillance and fully autonomous weapons use.

That means even if the Pentagon were to shift work away from Anthropic, it would face similar constraints with OpenAI regarding how AI systems can be deployed.

In an interview with CNBC on Friday, Altman said it was important for technology firms to work with the Pentagon “as long as it is going to comply with legal protections” and respect “the few red lines” common across much of the AI industry.

“For all the differences I have with Anthropic, I mostly trust them as a company, and I think they really do care about safety, and I’ve been happy that they’ve been supporting our war fighters,” Altman said. “I’m not sure where this is going to go.”

Anthropic’s Claude system was the first frontier AI model approved for use on the military’s classified systems. The Pentagon, however, recently gave the company a deadline to drop certain internal guardrails and allow its technology to be used for “all lawful use,” threatening to terminate a $200 million contract and designate the firm a “supply chain risk” if it refused.

Anthropic has said its concerns about autonomous weapons and mass surveillance stem from the technology’s current limitations and what it views as gaps in existing legal frameworks governing AI use.

Altman, in a memo to OpenAI staff first reported by The Wall Street Journal, said OpenAI was still working toward broader deployment of its models on classified systems and hoped the agreement could help “de-escalate” tensions between the Pentagon and Anthropic.

Last summer, OpenAI was among several AI companies to sign agreements with the Defense Department to “develop prototype frontier AI capabilities to address critical national security challenges in both warfighting and enterprise domains,” according to the Pentagon.

Until recently, Anthropic’s Claude had been the only model operating in classified military systems.

A Pentagon official told CNN this week that Elon Musk’s Grok, developed by xAI, is now approved for use in classified settings, while other companies, including OpenAI, were close to finalizing similar arrangements.

In his post, Altman said OpenAI is urging the Defense Department to offer the same terms to all AI providers. "We have expressed our strong desire to see things de-escalate away from legal and governmental actions and towards reasonable agreements," he wrote.

"We remain committed to serve all of humanity as best we can," Altman added, also noting, "The world is a complicated, messy, and sometimes dangerous place."