OpenAI has signed an agreement with the United States defence department to deploy its artificial intelligence models within the department’s classified network, marking a significant expansion of AI use in sensitive government systems.
Sam Altman, Chief Executive Officer, announced the deal on Saturday via his X account, describing it as the outcome of extensive discussions focused on safety, governance and responsible use of advanced AI technologies.
The agreement allows OpenAI’s models to operate within secure government infrastructure under strict guardrails and technical constraints.
Altman said the partnership aligns with OpenAI’s core safety principles, including prohibitions on domestic mass surveillance and maintaining human responsibility in the use of force, particularly in relation to autonomous weapon systems.
“AI safety and wide distribution of benefits are the core of our mission,” Altman said, adding that the defence department reflects these principles in its legal and policy frameworks and that they have been incorporated into the agreement.
OpenAI will deploy field engineers to support implementation and oversight, and the models will run exclusively on cloud-based networks rather than on local or edge infrastructure, a measure aimed at ensuring tighter operational control and security.
The development follows a series of policy moves in Washington targeting AI vendors working with federal agencies.
Under the administration of Donald Trump, federal agencies were recently directed to stop using software from AI firm Anthropic, while the Pentagon reportedly designated the company a supply-chain risk.
According to Bloomberg, OpenAI declined to comment on whether its new arrangement would replace services previously provided by Anthropic.
The agreement comes shortly after OpenAI closed a $110 billion funding round at a reported pre-money valuation of $730 billion, one of the largest capital raises in the technology sector.
The company said the funding will accelerate deployment of advanced AI tools across its platform and deepen strategic partnerships with firms including Amazon and NVIDIA.
OpenAI also disclosed strong product growth:
- Codex, its AI-powered coding assistant, has seen weekly users more than triple to 1.6 million since the start of the year.
- ChatGPT now serves over 900 million weekly active users, including more than 50 million paying subscribers.
- More than 9 million paying business users rely on ChatGPT for workplace applications.
The defence partnership signals growing integration of advanced AI systems into national security operations, even as debates continue globally over safeguards, oversight and the appropriate boundaries for military use of artificial intelligence.
