Department of Defense AI Strategy: Accelerating U.S. Military Artificial Intelligence Dominance

Key Chapter Title List

1. Accelerating U.S. Military AI Dominance
2. Acceleration Path
3. Seven Pioneer Projects
4. AI Computing Power
5. Data Access
6. Talent
7. Acceleration Expectations

Document Introduction

This document is an internal memorandum issued by the U.S. Secretary of Defense on January 9, 2026, titled "Department of Defense Artificial Intelligence Strategy: Accelerating U.S. Military AI Dominance." The memorandum aims to respond to the national policy outlined in President Trump's Executive Order 14179 regarding maintaining and enhancing U.S. global AI dominance. It clearly states that AI-enabled warfare and capability development will redefine the nature of military affairs within the next decade. The document emphasizes that this transformation, accelerated by AI innovation in the U.S. private sector, is a race. The Department of Defense must build upon the integrated technological advantage established during Trump's first term and accelerate progress to make warfighters more lethal and efficient, thereby constructing a combat force where every component, from the frontline to the rear, prioritizes AI.

To achieve this goal, the Department of Defense has established four strategic pillars: First, unleashing experimentation with leading AI models across the entire department and rewarding AI-first reimagining of traditional methods. Second, actively identifying and eliminating bureaucratic obstacles hindering deep integration. Third, focusing investments to leverage America's core asymmetric advantages in AI computing, model innovation, entrepreneurial vitality, capital markets, and the irreplicable operational data accumulated over the past two decades of military and intelligence operations. Fourth, executing a series of pioneer projects to demonstrate the accelerated execution speed, focus, and ethos required to stay ahead. These PSP projects also serve as vehicles for rapidly completing foundational enabling elements.

The memorandum details seven initial pioneer projects, categorized into three mission areas: operations, intelligence, and enterprise. In the operations domain, projects include Swarm Forge, Agent Network, and Terminator Foundry, aimed at exploring new human-machine collaborative tactics, developing operational management and decision-support agents, and accelerating AI-enabled simulation capabilities. In the intelligence domain, the Open Arsenal project aims to accelerate the process from technical intelligence to capability development; the Grant project is dedicated to shifting deterrence from a static posture to dynamic pressure with explainable outcomes. In the enterprise domain, the GenAI.mil project aims to provide direct access to the world's leading AI models for the department's three million military and civilian personnel; the Enterprise Agent project establishes a playbook for the rapid, secure development and deployment of AI agents to transform workflows. Each project has a single accountable leader and must conduct an initial demonstration to end-users within six months.

To ensure strategy implementation, the memorandum clarifies acceleration measures for three key enabling elements. Regarding AI computing power, the Department of Defense will invest significant resources to expand access to computing infrastructure from data centers to the edge, leveraging creative partnerships with leading companies to catalyze hundreds of billions of dollars in private sector investment. Regarding data access, the directive mandates enforcement of the DoD Data Decree, requiring all services and components to submit cross-security-level data catalogs to the Chief Digital and Artificial Intelligence Office within 30 days to further unlock data for AI development and mission advantage. Regarding talent, the directive requires the entire department to utilize special hiring and compensation authorities and tasks all components to submit AI recruitment and talent development plans within 60 days. The document concludes by emphasizing the core execution philosophy of speed wins and AI model parity, demanding the weaponization of learning speed and accepting that the risk of acting too slowly far outweighs the risk of imperfect coordination.