Skip to content

NSPM-11 — Artificial Intelligence in the National Security Enterprise

President Trump issued National Security Presidential Memorandum 11 (NSPM-11) on 5 June 2026. It is a binding presidential direction to the executive branch’s national-security enterprise, not legislation governing private AI development generally. It rescinds and replaces Biden-era NSM-25 and associated guidance.

NSPM-11 organises national-security AI policy around adoption, adaptation, assurance and accountability. It directs agencies to accelerate access to advanced commercial and open-source models from multiple suppliers, adapt systems to mission needs, preserve controllability and the constitutional chain of command, and avoid single-vendor dependency.

  • update DOD Directive 3000.09 on autonomy in weapon systems within 90 days and review it annually;
  • issue governance policy for AI use in national-security systems within 90 days, aligned with OMB M-25-21 where appropriate;
  • update procurement to onboard advanced models from multiple vendors and build a secure-compute roadmap;
  • establish an AI test range, private-sector security partnerships and joint red-team exercises;
  • create an AI National Security Strategic Reserve and a national-security AI curriculum;
  • develop joint risk-management, assurance and TEVV baselines within 120 days; and
  • permit contract termination, subject to limited waivers, for repeated conduct inconsistent with the memorandum’s policies.

EO 14409 creates a voluntary pre-release and cybersecurity interface for covered frontier models. NSPM-11 governs the government’s own defense and intelligence adoption, procurement, assurance and accountability architecture. Together they mark a security-focused qualification of the administration’s broader deregulatory posture, without creating a general model-licensing regime.