DOD Memo: Phase Anthropic Out of National Security Systems
Internal DOD directive (March 6), reported by CBS/The Hill: senior leaders ordered to identify and discontinue Anthropic models/services after supply-chain risk designation — timelines by criticality; reports link to Claude in sensitive ops including Iran campaign aspects. Anthropic: safeguards non-negotiable; lawsuits call designation retaliatory. Escalation after the label — not a civilian ChatGPT ban.
A supply-chain risk label is paperwork until someone orders uninstalls. The proprietary escalation: a March 6 internal DOD memo turning the Anthropic designation into operational removal from key national security systems.
Per CBS News and The Hill reporting on the memo addressed to senior military leaders and commanders: agencies must identify and discontinue Anthropic models and services; national security concerns tied to the company’s refusal to relax safeguards for certain military applications; implementation timelines vary by system criticality, with some immediate actions. Reports indicate references to Claude use in sensitive operations, including aspects of the Iran campaign. Anthropic maintains safeguards are essential; most commercial customers unaffected; lawsuits challenge the designation as retaliatory.
Claims vs checks
Memo existence/substance are secondary reporting of an internal directive (CBS, The Hill; also NYT/Reuters defense coverage). Anthropic position is company primary. Iran-campaign detail is as reported — not independently verified here.
Limits
- Internal memo scope can be narrower than headlines.
- Litigation outcome unknown at file date.
- Companion email/court and designation files cover cause chain.
Sources
- Internal DOD memo dated March 6, 2026 (obtained/reported by CBS News, The Hill).
- Anthropic news/statements (early March 2026).
- The New York Times, Reuters, and defense-focused publications.
- Related Senate letters and designation timeline reporting.