OpenAI is considering a contract to deploy its AI systems on NATO’s unclassified networks, extending the company’s recent move deeper into defense and government work just days after it struck a separate agreement with the Pentagon. Reporting on the discussions indicates the opportunity under consideration is limited to NATO’s unclassified systems, not classified networks.
The possible NATO deal would mark another step in OpenAI’s effort to build out its public-sector business. In June 2025, the company launched OpenAI for Government, a program focused on supplying advanced AI tools to U.S. public servants. On March 2, OpenAI published new details about its Pentagon agreement, saying it had reached a deal to deploy advanced AI systems in classified environments and outlining what it called three core red lines governing that work.
Those guardrails are central to the broader debate around military AI. OpenAI said its Defense Department agreement bars intentional use of its systems for domestic surveillance, autonomous weapons systems, and autonomous decision-making in national security contexts. The company also said the Pentagon had affirmed that OpenAI’s systems would not be used by agencies such as the NSA for domestic surveillance of U.S. persons.
OpenAI chief executive Sam Altman has been at the center of that shift. Reporting on the NATO talks indicates Altman told staff the company was looking at a NATO contract, before an OpenAI spokesperson clarified that the potential deployment would be for unclassified networks rather than classified ones. That clarification matters because it narrows the scope of the possible work and places it outside NATO’s most sensitive systems.
The NATO discussions are unfolding against the backdrop of a widening split inside the AI industry over defense work. Anthropic chief executive Dario Amodei has publicly said his company supports national security uses of AI but will not permit Claude to be used for fully autonomous weapons or mass domestic surveillance. Anthropic’s March 5 statement said the company had been designated a supply-chain risk by the Department of War after refusing to remove those limits, and it said it planned to challenge that action in court.
That dispute has turned OpenAI and Anthropic into visible symbols of two different approaches to government AI contracting. OpenAI has argued that its Pentagon deal includes more guardrails than prior classified AI agreements and has presented its approach as a workable framework for cooperation with defense agencies. Anthropic, by contrast, has framed its standoff with the Pentagon as a fight over whether the government can force private AI companies to abandon key safety restrictions.
President Donald Trump is part of the wider context around that fight. Coverage of Anthropic’s legal challenge says the Trump administration directed federal agencies to stop using Anthropic’s tools after the company’s clash with the Pentagon, intensifying the commercial stakes in the competition between major AI vendors for defense work. That has increased attention on which companies are willing to accept military contracts and on what terms.
A NATO arrangement would also broaden OpenAI’s defense footprint beyond the United States. NATO is a 32-member military alliance, and a deployment on its unclassified networks would still carry significant weight because it could place OpenAI’s systems into a large multinational institutional environment, even without touching classified data. Reports on the discussions indicate the talks remain exploratory rather than final.
The company’s earlier Pentagon deal already showed how closely cloud and infrastructure players could be tied to this market. OpenAI’s government work sits alongside longstanding defense relationships held by major technology companies including Microsoft and Amazon, both of which have deep existing ties to government and military cloud environments. That makes OpenAI’s latest moves part of a broader contest over who will supply the AI layer inside public-sector and defense systems.
With Sam Altman pushing OpenAI further into government AI deployments, Dario Amodei challenging the Pentagon’s demands from the outside, and Donald Trump’s administration reshaping the terms of the defense-tech relationship, the possible NATO deal has become more than a simple contract story. It is part of a larger fight over which companies will build military AI systems, where those systems can be used, and how much control developers will retain once governments come calling.
Read our coverage of Anthropic’s dispute with the Pentagon for another look at how AI companies are colliding with government power, defense policy, and model safeguards.
