The ChatGPT maker brings a former Pentagon chief aboard as its defense business moves from experiment to strategy
OpenAI has hired Mark Esper, the former US Secretary of Defense, to help expand its military AI work. The appointment was announced on November 2, 2026, and Esper is described as a strategic advisor for the company’s military AI initiatives.
A few years ago, OpenAI’s usage policies barred military and warfare applications outright. Now the company has a former Pentagon chief on its roster.
What OpenAI is building with the Pentagon
On February 9, 2026, the company announced that ChatGPT would become available on GenAI.mil, the military’s generative AI platform. The custom version is set to serve up to 3 million users.
Then, on February 27, 2026, OpenAI signed an agreement with the Department of Defense to deploy its AI models in classified environments.
The agreement also reportedly draws lines. It sets boundaries against mass domestic surveillance and against the use of autonomous weapons.
On top of that, OpenAI has gathered unclassified defense contracts valued at up to $200 million.
From prohibition to partnership
The pivot started in 2024, when OpenAI removed the prohibitions on military and warfare uses from its policies. The policy shift cleared the way for the GenAI.mil rollout, the classified deployment deal and the contract pipeline that followed.
What this means for OpenAI and the defense AI market
The $200 million in unclassified contracts is the number to watch. If the classified deployment agreement and Esper’s hiring translate into larger awards, OpenAI’s defense revenue could grow well beyond that ceiling.
OpenAI’s stated limits on domestic surveillance and autonomous weapons will face pressure as its government work deepens. Boundaries written into an agreement are only as durable as the enforcement and transparency behind them.
