IC Technology & Surveillance — 2026-04-29

Pentagon Signs Classified AI Deal with Google After Anthropic Blacklisting

Google signed an agreement allowing the Pentagon to use its Gemini AI for classified military work, making it the latest firm alongside OpenAI and xAI to support classified operations. The deal comes after DOD blacklisted Anthropic's Mythos model and reflects the Pentagon AI chief's stated need to diversify AI vendors. Over 700 Google employees signed a letter opposing the arrangement.

Analysis
The DOD's multi-vendor AI strategy, now spanning Google, OpenAI, and xAI for classified work, reflects deliberate insulation against single-provider risk, with Anthropic's Mythos exclusion signaling that vetting authority is real, not ceremonial. Per a single TechCrunch report, at least one additional major AI firm will likely sign a classified defense contract by end of 2026, driven by the Pentagon AI chief's stated diversification mandate and accelerating vendor onboarding. The stronger constraint on that expansion may instead be internal labor dynamics rather than inter-agency policy. Over 700 Google employees signed in opposition, a divergence from hardware and aerospace precedents where employee dissent carried no comparable structural leverage.

2 sources