Building the Full AI Stack as the Core of Contemporary U.S. Science, Technology, and Innovation Policy

Abstract

Introduction. Donald Trump’s second presidential term in 2025 has brought about a radical transformation of U.S. science, technology, and innovation (STI) policy. The administration has abandoned the green-energy priorities of its predecessor and has repositioned artificial intelligence (AI) as the central system-forming technology of national development.Problem Statement. The new strategy is centered on the AI “full stack,” understood as a fully American technological chain encompassing semiconductors, data centers, computing infrastructure, and soft - ware platforms. This strategy is being pursued amid intensifying technological competition between the United States and China, as well as growing concerns regarding the emergence of an investment bubble in the AI sector.Purpose. The purpose of this study is to analyze the transformation of U.S. STI policy during 2024—2025, with particular emphasis on the formation of the AI “full stack,” and to assess the economic and institutional consequences of these changes in the context of global technological competition.Materials and Methods. The study is based on an analytical review of official documents of the U.S.federal government, reports issued by leading research centers, including Stanford Human-Centered Artificial Intelligence, RAND Corporation, and Atlantic Council, as well as statistical data and specialized scholarly and analytical publications.Results. Contemporary U.S. STI policy is concentrated around five principal directions: deregulation of AI technologies; semiconductor reindustrialization through protectionist tariff policies; large-scale construction of data centers (including the Stargate initiative valued at USD 500 billion); expansion of energy-generation capacity, ranging from nuclear energy to coal-based power generation; and strategic technological confrontation with China through restrictions on exports of advanced semiconductors. A key finding of the study is the fundamental distinction between AI and previous hegemonic technologies, such as semiconductors and Internet protocols of the 1990s. Unlike these technologies, AI functions as a knowledge infrastructure that inherently requires openness, which directly contradicts the logic of monopolization. The study has also identified significant risks of a potential crisis in the AI industry during 2026—2027.Conclusions. The contemporary U.S. STI strategy creates both challenges and opportunities for Ukraine. An optimal strategic approach involves positioning Ukraine as an important component of the American AI stack, particularly as a unique testing environment for distributed defense-oriented AI solutions, while simultaneously leveragingopen-source AI technologies and preparing for the increasing role of public investment in the event of a cooling of private capital markets.