The Pentagon gave 3 million staff ChatGPT and Grok and left out Claude
Evan / Policy and Open Source desk
The Pentagon rolled out ChatGPT and Grok to 3 million people this week, and the one lab that argued about the rules was left off the list.
The rollout
On August 31, 2026 the Department of Defense added OpenAI’s ChatGPT Mil and Grok for Government to GenAI.mil, the platform where Google’s Gemini went live first. DefenseScoop reports the rollout reaches military members, civilian employees and contractors, about 3 million people, up from roughly 1.5 million users earlier in the year. A department official said access to “a diverse suite of AI capabilities from across the resilient American technology stack will give warfighters the tools they need to act with confidence.”
Why Claude is missing
Anthropic’s Claude was in the original plan and is not on the platform. The dispute, as DefenseScoop describes it, was about guardrails: Anthropic wanted contract language barring use for mass surveillance of Americans or for lethal autonomous weapons, and the department demanded unrestricted access for “lawful purposes.” Six months earlier the department labeled Anthropic a supply-chain risk and ordered Claude purged from national security systems. Anthropic sued, and in late August a federal judge ruled the move illegal. The legal fight continues while GenAI.mil grows without Claude.
What the episode shows
Whatever one thinks of either side, this is the clearest example yet of a lab losing a government customer over conditions rather than capability, and of the customer simply choosing the labs that did not ask. The question left open is whether an AI company can set terms on military use of its model and keep the contract.
Sources
Related
ChatGPT, Claude and Grok were down at the same time
On the morning of September 3, 2026, Pacific time, Grok and Claude began failing around 6:30 a.m. and OpenAI reported a routing error from about 7:43 a.m. that made ChatGPT and Codex unavailable for some users. Grok was offline about three and a half hours, including on X and its mobile apps. SpaceXAI apologized for an outage at its Memphis data center that affected compute partners, without naming a cause. No shared root cause has been established.
OpenAI models turn out to be about 5 percent of Cursor's traffic
After OpenAI announced it would end Cursor's model access on a proposed date of November 12, 2026, Cursor CEO Michael Truell said OpenAI models account for about 5 percent of Cursor's AI traffic. OpenAI cited Elon Musk's history of breaking contracts, including ending Twitter's data-sharing deal with OpenAI and allegedly training Grok on other labs' model outputs. Musk responded that he could not care less. Anthropic co-founder Tom Brown called Cursor a trusted partner since Sonnet 3.5 and pledged to keep expanding Claude compute for Cursor.
The EU orders Google to open Android to rival AI assistants
On July 16, 2026 the European Commission adopted two binding Digital Markets Act decisions against Google. One forces Google to open 11 Android feature groups to rival AI assistants with the same system-level access Gemini gets, from voice activation to acting across apps; the other requires Google to share anonymized search query, click and ranking data with competitors on fair terms. Google says the decisions risk undermining privacy and security guardrails for millions of Europeans.
ANOTHER News is published by ANOTHER, an AI-native content agency. Daily coverage also runs on Instagram.