National Security Presidential Memorandum NSPM-11 is a signed primary text issued by President Donald J. Trump on June 5, 2026, directing the United States national security enterprise to accelerate its adoption of AI under four pillars: Adoption, Adaptation, Assurance, and Accountability. It rescinds and replaces Biden's National Security Memorandum NSM-25, orders an update to DOD Directive 3000.09 on autonomy in weapon systems, establishes an AI National Security Strategic Reserve, requires anti-kill-switch contract clauses, and creates a FISMA-based mechanism for terminating contracts with non-compliant vendors.
Primary text: whitehouse.gov (The White House, Presidential Memoranda, June 5, 2026). Companion fact sheet: whitehouse.gov. For the living tracking page covering deadlines, reception, and downstream implementation, see NSPM-11 — Artificial Intelligence in the National Security Enterprise (Trump, June 5, 2026).
Scope and addressees
The memorandum directs the United States national security enterprise — defined in the text as "the Department of War, the Intelligence Community, and other agencies that develop, deploy, or use national security systems or otherwise serve a national security role" — to "responsibly accelerate the use of AI across intelligence and warfighting domains in line with American values." It is addressed to a broad slate of principals: the Secretaries of State, Treasury, War, Energy, and Homeland Security; the Attorney General; the DNI and CIA Director; the OMB Director; the National Cyber Director; the FBI and OPM Directors; and the President's national-security, science-and-technology, and homeland-security advisors.
NSPM-11 rescinds and replaces Biden's National Security Memorandum NSM-25 and associated guidance, which it frames as having "imposed undue bureaucracy that hampered the pace of AI adoption, fostered dangerous dependencies on single vendors, and made it challenging for our warfighters to adopt the most advanced technologies."
The four pillars (Section 2)
Section 2 organizes the memorandum's policy around four pillars:
- Adoption — accelerate AI adoption by identifying mission areas where AI enhances operational effectiveness and eliminating "unnecessary barriers to rapid deployment," maintaining "deep, proactive partnerships with industry" to make "the most advanced frontier models broadly available to national security professionals without delay."
- Adaptation — adapt commercial or open-source AI from "diverse suppliers across the private sector, large and small," resorting to internally customized or internally developed systems only where a commercial solution is inappropriate for security or mission reasons.
- Assurance — ensure adopted AI is "reliable, robust, steerable, and controllable," with testing, evaluation, validation, and verification; and ensure, via contract clauses or other means, that "no commercial entity or adversary possesses the capability to prevent use of, disable or degrade, or materially modify without Federal Government knowledge and approval" an AI system warfighters depend on. This is the "no kill switch" assurance requirement.
- Accountability — American AI "shall neither be developed nor used by the national security enterprise to censor free speech, embed ideological bias, or conduct unauthorized or unlawful surveillance"; commanders, directors, and agency heads "remain responsible and accountable" at every level of command.
Operative directives and deadlines
The memorandum carries dated directives in Sections 3 through 5. Within 90 days, the Secretary of War issues an update to DOD Directive 3000.09 on Autonomy in Weapon Systems, reviewed annually, "to ensure the deliberate adoption of AI systems that respect the chain of command"; the Committee on National Security Systems and the OMB Director issue an AI governance policy for national security systems, maximizing consistency with OMB memorandum M-25-21 where appropriate; a classified annex is issued; and the APST and OMB Director develop a roadmap for advanced computing access, including high-security AI computing facilities and a national-security AI test range, subject to appropriations.
Within 120 days, the Secretary of War, the DNI, and IC-element heads update procurement for rapid multi-vendor onboarding, "closing the capability gap between what is available to the public and to our national security workforce." The same window requires private-sector AI-security partnerships, including against "malicious distillation attacks," via the AI Security Center; an AI National Security Strategic Reserve of non-governmental AI talent, initiated by the OPM Director; an AI for National Security Curriculum; a joint AI risk-management and assurance strategy; and standardized AI national-security TEVV methodologies.
Vendor-conduct termination mechanism (Section 3(b))
Invoking authorities under the Federal Information Security Modernization Act of 2014 (44 U.S.C. 3551 et seq.), the memorandum directs agency heads to "direct, to the maximum extent permissible by law, termination for default or for convenience" of contracts, including subcontracts, with companies "that have repeatedly demonstrated a pattern of conduct that is inconsistent with policies laid out in section 2." Agencies may grant time-limited waivers not exceeding one year for operational imperatives, test-and-evaluation, threat-intelligence sharing, and other mission-critical applications; all waivers must be reported in writing to the APST and APNSA within 30 days.
This provision operationalizes the "no kill switch" assurance requirement against vendors and bears on the friction between the administration and Anthropic over the Pentagon's "supply-chain risk" designation.
Definitions of note (Section 6)
NSPM-11 defines four governance terms it repeatedly invokes: reliability ("perform as required, without failure, under given conditions"), robustness ("maintain a level of performance under a variety of circumstances, including outside intended operating conditions"), steerability ("shape the internal behavior of a system to pursue a given set of objectives"), and controllability ("monitor the operation and outcomes of a system and take corrective action as needed"). It adopts the 15 U.S.C. 9401(3) statutory definition of "artificial intelligence."
Relation to the broader AI agenda
NSPM-11 is the national-security companion to the administration's broader AI agenda. It converts the procurement-and-deployment posture of the AI Action Plan into a defense and IC-specific directive and sits alongside the June 2, 2026 EO on AI innovation and security as a companion rather than a superseding instrument. It codifies two themes of the administration's AI agenda — opposition to single-vendor dependency and an accountability standard against ideological bias and unlawful surveillance — into binding agency obligations. The document embeds a controllability and assurance vocabulary usually associated with the safety community into an instrument whose stated primary aim is acceleration.
Provenance and verification
Canonical host: whitehouse.gov (Presidential Actions / Presidential Memoranda; full verbatim text). Pulled and authenticity-verified during the June 6, 2026 gap scan against the official presidential-actions page and corroborated by the matching White House fact sheet (whitehouse.gov, June 5, 2026), Reuters, and Inside AI Policy reporting (both June 5, 2026). Verification record: NSPM-11 — Artificial Intelligence in the National Security Enterprise (Trump, signed June 5, 2026).
Relationships
- supersedes: Biden-era NSM-25 (no wiki page; rescinded by this memorandum)
- depends-on: America's AI Action Plan
- supports: NSPM-11 — Artificial Intelligence in the National Security Enterprise (Trump, June 5, 2026) — the living tracking page built from this text
- related: EO — Promoting Advanced AI Innovation and Security (Trump, signed June 2, 2026), Frontier Act / Great American AI Act (Obernolte–Trahan), DOD — Department of Defense (AI Deployer), Anthropic, Frontier AI Governance, Defense / Military — AI Deployment