Pentagon adds ChatGPT and Grok to GenAI.mil, still no Claude
A federal judge just ruled Anthropic's 'supply-chain risk' label unlawful retaliation — yet the DoD's 3M-user AI portal still runs everyone but Claude.
Copy markdown
ChatGPT and Grok clear IL5 for 3M troops
The Pentagon's GenAI.mil portal added OpenAI's ChatGPT and xAI's Grok (via Starshield AI) on Aug 31, joining Google's Gemini — all three cleared at Impact Level 5 for sensitive unclassified data across 3M+ members, civilians, and contractors. It's the widest frontier-model rollout in the US government yet, and Claude is the one name missing.
A judge just voided Claude's 'supply-chain risk' label
On Aug 28, District Judge Rita Lin ruled the Pentagon's blacklisting of Anthropic was unlawful First Amendment retaliation, calling its national-security evidence 'slim.' The legal barrier is gone on paper — but the Aug 31 expansion still shipped without Claude, and the DoD has not reinstated it.
The fight is over surveillance and autonomous weapons
The rift traces to Anthropic's refusal to drop usage-policy clauses barring Claude from mass surveillance and fully autonomous lethal weapons; the Pentagon demanded 'unrestricted access' for lawful use. The Register reported the original blacklist even cited Claude capabilities it did not have.
What actually changes for your stack
Nothing for commercial Claude today — the API, Claude Code, and consumer plans are untouched. The signal for builders: Anthropic will defend its acceptable-use limits even against the DoD, so don't count on surveillance or weapons carve-outs appearing in the terms you build against.