NSPM-11

A New Directive for National Security AI

The Four Pillars

The memorandum is built on four pillars. Adoption directs agencies to eliminate bureaucratic barriers and get frontier AI models into the hands of national security professionals without delay. Adaptation tells agencies to lean on commercial and open-source AI from a diverse set of private-sector suppliers, building custom solutions only when the mission truly demands it. Assurance requires that AI systems be reliable, robust, steerable, and controllable — and protected from interference by adversaries or vendors. Accountability prohibits the use of AI to censor speech, embed ideological bias, or conduct unauthorized surveillance.

Short Timelines, Bold Mandates

The timelines are short. Within 90 days, agencies must update the Pentagon's directive on autonomy in weapon systems and issue new AI governance policy for national security systems. Within 120 days, they must overhaul AI procurement processes, stand up a national security AI test range, launch an AI training curriculum, deliver a joint AI risk-management strategy, and establish an "AI National Security Strategic Reserve" of private-sector talent that can be called on to support federal AI efforts.

What It Means for Industry

For companies serving the national security market, the takeaway is clear: the government is moving from cautious evaluation of AI to rapid, large-scale procurement — with new compliance expectations around security, testing, and vendor accountability attached. Organizations that understand these requirements early will be best positioned to compete for the wave of contracting activity NSPM-11 sets in motion.