Anthropic AI models used in US Iran strikes despite ban
Get the Tech newsletter
Daily tech — startups, AI labs, chips, the launches that shape the next decade. Free.
- President Trump directed civilian agencies to stop using Anthropic products, while Anthropic was given six months to wind down its Department of Defense work.
- U.S. and Israel launched a surprise attack on Tehran, beginning a conflict before Trump's directive could be fully implemented.
- Anthropic models are being used for real‑time targeting and target prioritization in the U.S. aerial campaign against Iran, according to The Washington Post.
- Anthropic and Palantir's Maven system together suggested hundreds of targets, provided precise coordinates, and ranked targets by importance for Pentagon strike planning.
- Pete Hegseth pledged to label Anthropic as a supply‑chain risk, yet no official designation has been made, leaving no legal barrier to its use.
- Lockheed Martin and other defense contractors have begun replacing Anthropic’s models with rival AI services, per Reuters.
- J2 Ventures reported that ten of its portfolio companies have stopped using Claude for defense applications and are actively seeking replacements.
Why it matters: The Pentagon continues to rely on Anthropic's AI for live targeting, giving the U.S. a tactical edge in the Iran strike, while Anthropic loses defense contracts and faces a potential supply‑chain risk label that could force a costly migration to rival models for its customers.
