Pentagon pulls plug on Claude after Anthropic refused to lift limits on surveillance, weapons
mark_k · x · 2026-10-07
- Citing an official speaking to the BBC, the Pentagon has officially stopped using Anthropic's Claude. Anthropic was blacklisted in February with an August-end deadline to stop using its tools, yet sources say Claude was still in use as recently as last week, including in military operations against Iran.
- The split followed Anthropic's refusal to remove restrictions over concerns about mass surveillance and autonomous weapons.
- The Pentagon has since signed contracts with Google, xAI and OpenAI, but Claude was deeply embedded in Palantir's Maven intelligence platform, making removal far harder than cancelling a subscription.
More from Companies & People
- Anthropic design team case study: code as clay and design at AI speed — joellewenstein · 2026-10-07
- If AI Handles Most Support Tickets, the Support Org Reorganizes Into Four New Roles — Informal-Dust4499 · 2026-10-07
- MiniMax lines up four SF Tech Week events on models, inference and agents — MiniMax_AI · 2026-10-07
- Sergey Karayev: Claude is autumn-coded humanpunk, rivals are winter cyberpunk — sergeykarayev · 2026-10-07
- Atlassian and OpenAI team up to ground frontier models in enterprise context via Teamwork Graph — davidhoang · 2026-10-07
- Consultant lands six-figure enterprise AI deal with just context layering and security — verrsane · 2026-10-07