- Federal AI R&D funding is difficult to track because the government lacks standardized, timely, and accurate reporting on authorizations and appropriations, especially during continuing resolutions and irregular appropriations cycles.
- The report states that the United States spends roughly $3 billion on AI R&D, far below the $32 billion target set by the National Security Commission on Artificial Intelligence.
- The report argues that increased AI R&D spending should prioritize foundational research and critical infrastructure across the AI stack, including talent, data, hardware, algorithms, applications, and integration.
- Five organizations are identified as leading federal AI R&D investors: the Department of Energy, National Science Foundation, Department of Commerce, National Institutes of Health, and NASA.
- Several Department of Energy initiatives, including Advanced Scientific Computing Research, Microelectronics Science Research Centers, and FASST, are described as underfunded or not fully funded relative to their intended roles or authorizations.
This September 2024 report analyzes federal funding for artificial intelligence research and development over the previous five years, drawing on department and agency budget requests and other cited sources. It emphasizes that tracking AI R&D spending is difficult because reporting on authorizations and appropriations is not standardized, timely, or consistently accurate, and because continuing resolutions and irregular appropriations cycles obscure program-level planning. The report argues that U.S. non-defense AI R&D investment remains far below the National Security Commission on Artificial Intelligence’s $32 billion benchmark, with current spending described as roughly $3 billion. It focuses on five leading federal AI R&D organizations: the Department of Energy, National Science Foundation, Department of Commerce, National Institutes of Health, and NASA, while also noting selected programs at other agencies. Across these institutions, the report finds that major AI, microelectronics, computing, standards, biomedical, and data programs are stagnant, under-resourced, or funded below authorized levels. It recommends strategically increasing federal non-defense AI R&D spending by building on effective programs, filling gaps neglected by private capital, and pursuing moonshot-style goals that advance the AI stack. The report frames these choices as central to U.S. innovation, national security, and technological leadership amid global competition.
Increasing federal spending for non-defense AI R&D to reach $32 billion by 2026 is not merely an ambitious goal. It is a strategic imperative.
SCSP
- Strategically increase federal non-defense AI R&D spending. This increased funding should build upon successful existing programs, fill critical gaps, and set audacious moonshot goals.
- Build upon successful existing programs. Existing, effective programs like NSF's AI research and related activities can be empowered, expanded, and properly resourced.
- Fill critical gaps in research and development. Federal investments should target areas crucial for future technological advancement but may be overlooked by private sector funding.
- Set audacious goals and embrace a moonshot mindset. The United States government should fund ambitious, AI-enabled, whole-of-nation efforts to push the boundaries of what is technologically possible.
Published by the Special Competitive Studies Project. Co-publishers and external authors, where they exist, are credited in the byline above.