The Trump administration’s new technology strategy places drones, autonomy and other emerging military systems at the center of its approach to deterring China, while directing national-security research toward large foundation models. The document also calls for broader experimentation and new pathways for smaller defense companies to enter federal markets.
The strategy identifies undersea systems, space, and artificial intelligence and autonomy as priority areas for military research and spending. Within those areas, it lists multi-agent systems and swarm intelligence, along with applications of autonomy in robotics and command and control, as critical technologies for the military.
It also calls for combinations of lower-cost, sometimes lower-tech or attritable platforms that can operate in larger numbers alongside a smaller number of sophisticated systems. The approach would include drones and drone swarms working with manned aircraft and legacy platforms.
The emphasis follows uneven efforts over the past decade to field autonomous aircraft alongside fighter jets and ships. Those efforts have faced constraints from internal acquisition policies and earlier commitments to other programs, according to Defense One. The Replicator program, which was presented as a pathfinder for deploying large numbers of lower-cost systems, received $1 billion in funding.
The strategy seeks to alter the way the Pentagon buys emerging technology. It urges officials to use a mix of contract types, including performance-based contracts and other transaction authorities, rather than relying only on traditional acquisition approaches. It also endorses faster, nontraditional methods that may involve greater risk in exchange for the possibility of higher rewards.
The document extends its support for new defense companies beyond Pentagon procurement. It calls for changes to foreign military sales rules and export controls that have hindered sales of some defense technologies to U.S. allies. Defense One reported that such changes could help newer companies attract early investment and remain viable while developing their products.
The strategy also asks federal agencies to create streamlined pathways for smaller firms. Those pathways could include Cooperative Research and Development Agreements and Agreements for Commercializing Technology, which may give companies access to federal infrastructure and other forms of collaboration.
Its AI provisions take a more specific approach. The strategy names foundation models, including large language, multimodal and world models, among the critical technologies supporting national security. It does not name open-weight models or open-source software development.
That choice excludes approaches that use freely available model parameters with alternative architectures designed to reduce computing and energy demands. Defense One reported that some of these systems can approach the performance of leading commercial models while requiring fewer resources.
The distinction has operational implications for military users. Frontier AI systems can depend on extensive cloud infrastructure, data access and reliable connectivity, conditions that may not be available to deployed units. Defense One cited a Carnegie Endowment for International Peace report saying that Ukraine’s combination of Western cloud access, domestic data centers and forward-deployed computing nodes was already under strain as it expanded AI use for targeting and coordination.
Michael Schiffer, a partner at Scalare Advisors and a former deputy assistant secretary of defense for East Asia, argued that the strategy’s treatment of AI could benefit China. The strategy’s focus on foundation models favors an industry structure dominated by a small number of companies with access to large amounts of computing infrastructure, while leaving other technical approaches outside its named critical-technology priorities.
The strategy therefore combines an expansive view of military autonomy with a narrower definition of the AI technologies it highlights. Its procurement provisions and support for smaller defense firms are intended to widen experimentation and access to federal markets. Its AI framework, however, centers on foundation models and does not give comparable recognition to open-weight or open-source alternatives.
Whether the approach broadens the defense technology base will depend on how the priorities and acquisition changes are applied. The document establishes areas for research, preferred contracting methods and possible routes for smaller companies to work with the government, but it does not specify how quickly those measures will produce fielded capabilities. Defense One