Photo: Air Force Staff Sgt. John Wright, DOD (Public domain), via Wikimedia Commons (source)
The Defense Department's enterprise generative AI platform, GenAI.mil, drew more than 2 million users in a single week this month, the department's top AI official said Wednesday, Sept. 23, as another senior official said the department plans to bring frontier models onto its classified networks.
"As of last week, we've had over 2 million people in a week across all three models log in and use AI," Cameron Stanley said in a discussion with Punchbowl News, Defense One reported. The platform, launched in December, offers three models for unclassified work: Gemini for Government from Google, OpenAI's ChatGPT Mil and xAI's Grok for Government. About 3 million military, civilian and contractor personnel can access it.
Growth since launch
Stanley said about 80,000 people across the department were using AI before GenAI.mil arrived, and more than 300,000 logged on in its first week. After Google's Agent Designer tool was added in March, he said, users built more than 50,000 agents "within two weeks," mostly simple helpers for daily tasks. In July, he had put the platform's record at 1.7 million users and more than 100,000 custom agents, according to Defense One.
A day earlier, at the DefenseScoop-hosted DefenseTalks conference, James Mazol, deputy under secretary of defense for research and engineering, said 1.7 million of the department's 3 million personnel have used the platform. "Of that 1.7 [million], I would say we've got about half a million power users," he said, describing people who use generative AI nearly every day, DefenseScoop reported. The Army, Navy, Marine Corps, Air Force and Space Force have adopted it as their preferred enterprise AI system.
Classified networks and compute
Mazol said the department will "proliferate frontier models" beyond its controlled unclassified network to SIPR, the secret network, to JWICS for top-secret work and into special access programs. He did not give a timeline in the remarks DefenseScoop reported.
Stanley flagged computing capacity as a constraint. "AI is getting put into everything. That scares me, mainly because of capacity," he said at DefenseTalks, adding that compute shortages affect industry broadly. He said the department needs models sized "for what we're using them for" to cut inference costs.
Asked about calls to slow AI development, Stanley told Punchbowl that "unilateral disarmament is the wrong approach," according to Defense One. President Trump said at the United Nations on Tuesday that the U.S. would "encourage it, not rein it in."
What's next
Stanley said the department is reworking how it integrates AI agents and maps agent-based workflows, DefenseScoop reported. No date was given for the first frontier models on classified networks.
Sources
- Defense One: GenAI.mil saw more than 2 million users in one week, top DOD official says (2026-09-23)
- DefenseScoop: GenAI.mil attracts about half a million ‘power users’ as Pentagon pushes forward with frontier models (2026-09-23)
Motion Media News corrects errors promptly. To report one, email hello@motionnews.studio.
Added to the Motion Media archive on Sep. 28, 2026.
