The AI platform of the Pentagon surged from 80,000 users to 1.5 million within a span of six months.

      TL;DR: GenAI.mil has reached 1.5 million daily users, a significant increase from 80,000 at its launch, following the Pentagon's implementation of Google Gemini and clarification of AI usage policies.

      According to the chief technology officer of the Department of Defense, the generative AI platform GenAI.mil now boasts 1.5 million daily users within the DoD, which accounts for nearly half of its 3.5 million workforce. Six months prior, the user count was below 100,000. Emil Michael, the undersecretary of defense for research and engineering, shared these figures at a Hudson Institute event last week. At its launch in December 2025, only 80,000 personnel were utilizing the platform. Michael noted that the initial low usage stemmed from uncertainty about accessing the tool, its applications, and the associated guidelines.

      “It wasn’t really clear where to go for it, what you could use it for, the rules were unclear, so we just blew through that,” Michael stated.

      The launch of Google’s Gemini on the Pentagon’s unclassified networks marked a pivotal moment, resulting in a surge in daily usage. The Pentagon has since incorporated OpenAI’s ChatGPT and xAI’s Grok into the platform, offering all 3.5 million DoD employees access to various commercial AI models via a single portal.

      Most AI applications among personnel are administrative, with workers using it for tasks such as drafting job descriptions, summarizing meeting notes, and creating budgets. A more significant use case involves congressional reporting: work that used to take 200 hours can now be completed in five by inputting source documents into the AI to generate the report.

      “More and more people are like, ‘Oh my God, I could write a job description,’” Michael remarked. “I mean, very simple things to more exquisite things.”

      The rapid adoption rate is noteworthy, not for its novelty but for its pace. The Pentagon is experiencing the same productivity benefits seen in corporate sectors since large language models became widespread. Michael acknowledged that they are simply trying to catch up to what has become standard in the commercial realm.

      The platform has evolved beyond basic chatbot functions. In April, the Pentagon reported that DoD personnel had developed over 100,000 semi-autonomous AI agents using the Gemini’s Agent Designer tool in under five weeks. These agents operate at Impact Level 5, the highest tier for unclassified sensitive data, undertaking tasks such as drafting after-action reports, analyzing operational data, and reviewing images.

      The financial commitment corresponds with the adoption scale, with the fiscal 2027 defense budget requesting $54.6 billion for the Defense Autonomous Warfare Group, significantly up from the previous year's AI and autonomy budget of $13.4 billion. The Pentagon’s overarching stance is that AI should be integrated across every function, from administrative tasks to battlefield decision-making.

      This ambition has led to some tension. Over 580 Google employees, including senior DeepMind researchers, signed a letter in April urging CEO Sundar Pichai to decline classified military AI work for the Pentagon, citing concerns that Google wouldn’t be able to oversee the use of its AI on air-gapped networks. This letter followed the deployment of Gemini to the Pentagon’s unclassified workforce and ongoing talks about classified access under “all lawful uses” terms.

      The Pentagon has since established classified AI agreements with seven companies – including Google, Microsoft, Nvidia, and OpenAI – aimed at replacing the safety regulations that Anthropic insisted on before being excluded from Pentagon contracts. The message to AI firms is clear: the military will determine usage parameters, not the vendors.

      However, the unclassified use of GenAI.mil presents a different scenario. Drafting job descriptions and congressional reports are examples of tasks where AI’s limitations, such as hallucination, factual inaccuracies, and confidently incorrect summaries, pose relatively low risks. Errors in a draft report can typically be caught by human reviewers. The Pentagon has not released data on error rates, accuracy metrics, or any internal evaluations of GenAI.mil's output quality.

      Michael characterized the growth as organic. The Pentagon shared case studies demonstrating how employees utilized the platform, disseminating this information throughout the department. He noted that exposure to consumer AI tools outside of work also aided adoption, as personnel were already acquainted with how chatbots function.

      Currently, five out of six military branches have identified GenAI.mil as their main enterprise AI platform. The jump from 80,000 to 1.5 million users in six months represents the fastest enterprise AI rollout in government history. Whether the quality of the output justifies the swift adoption is a question the Pentagon has yet to address publicly.