The Pentagon has launched military-tailored versions of ChatGPT and Grok on its GenAI.mil portal, giving 3 million civilian and military personnel access to custom frontier AI tools built around what the Department of Defense calls "warfighter needs." ChatGPT Mil comes out of OpenAI's OpenAI for Government program, while Grok for Government ships under SpaceX's Starshield AI, the secure satellite network built on Starlink infrastructure. Both are now live inside the same portal that debuted last year with Google Gemini.
Uptake is already substantial. GenAI.mil has onboarded 1.7 million unique users, roughly 57% of the Department of Defense's 3 million-person workforce, before the ChatGPT and Grok additions arrived. The portal exists specifically so DoD employees can use commercial frontier models without routing sensitive government data through consumer channels, and the military versions are exempt from the data collection baked into ordinary consumer products.
ChatGPT Mil will feel familiar to consumer ChatGPT users. The Defense Department says it will support chat, files, projects, and custom GPTs, with a focus on document-heavy unclassified work: administrative tasks, logistics, planning, and policy drafting. More features are slated to arrive over time.
Key facts
- 01The Pentagon launched ChatGPT Mil and Grok for Government on the GenAI.mil portal, opening access to 3 million military and civilian personnel.
- 02GenAI.mil has already onboarded 1.7 million unique users, more than half of the Department of Defense's 3 million-person workforce.
- 03Anthropic's Claude is absent from the portal after the Trump administration labeled the company a supply-chain risk, a designation Anthropic is contesting in court.
- 04Grok for Government ships under SpaceX's Starshield AI, the secure satellite network built on Starlink infrastructure.
- 05The Pentagon has separately signed AI deals with Amazon Web Services, Microsoft, Nvidia, and Reflection AI.
Grok for Government is being positioned in more militaristic terms. The department says Starshield AI's Grok for Government will deliver "immediate productivity gains, stronger knowledge continuity and more secure and efficient collaboration," and will let military users "execute missions faster and with greater precision across numerous operational contexts, ranging from market research analysis for acquisition professionals to supply-chain management for logisticians." The framing signals a broader operational surface than ChatGPT Mil's back-office pitch.
The most conspicuous absence is Claude. Anthropic's model is not on the portal after the company refused to give the Pentagon unrestricted use of its AI tools and instead insisted on retaining certain safety guardrails. The Trump administration subsequently designated Anthropic a supply-chain risk — a label Anthropic is currently fighting in court — and the Pentagon has redirected its procurement toward vendors willing to operate on the department's terms.
That redirection has been broad. The Pentagon has struck AI deals with Amazon Web Services, Microsoft, Nvidia, and Reflection AI alongside the OpenAI and xAI additions announced on August 31, 2026. The pattern suggests the department is deliberately building a multi-vendor stack rather than standardizing on a single frontier lab, which reduces dependence on any one provider's policy stance.
For OpenAI, ChatGPT Mil is a meaningful expansion of a government footprint the company has been building through its OpenAI for Government track. Getting inside a portal used by more than half of DoD personnel puts the product in front of an installed base that few enterprise sales cycles could match, and the unclassified-work framing keeps the deployment inside a compliance envelope OpenAI has already navigated.
For xAI, the Starshield AI packaging is arguably the more strategic win. Bundling Grok with SpaceX's secure satellite network ties Elon Musk's AI company to defense infrastructure his other company already operates, and gives xAI a distribution path that OpenAI, Microsoft, and Google cannot easily replicate. It also positions Grok as the model of choice for operational rather than administrative use cases, at least in the department's public framing.
The obvious caveat is that portal access is not the same as mission-critical adoption. Onboarding 1.7 million users tells you the login page works; it does not tell you how many of those users rely on the tools daily, what workflows have actually been migrated, or how the department will evaluate output quality across ChatGPT Mil, Grok for Government, and Gemini when they sit side by side. Nor is it clear how the Anthropic supply-chain-risk designation resolves — a court ruling in Anthropic's favor would reopen the door to Claude and complicate the current vendor mix.
The Pentagon's move locks in a template other federal agencies are likely to copy: a centralized portal, multiple frontier vendors, and hardened data terms that exclude any lab unwilling to meet them. That is a durable tailwind for OpenAI and xAI and a concrete cost for Anthropic, whose safety-guardrail stance is now measurably narrowing its federal addressable market. Whichever way the court case lands, the message to frontier labs from the largest government AI buyer in the world is that terms of service are now a competitive weapon.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
