Key Takeaways
- NSF invests $700M+ annually in AI research — the largest non-defense funder of foundational AI in the U.S.
- 25+ National AI Research Institutes funded at $16–20M each — new competitions open periodically; watch NSF.gov for announcements
- Individual researchers can access AI funding through CAREER awards, standard CISE research grants, and cross-directorate programs
- ExpandAI — up to $1M for Minority-Serving Institutions to participate in NSF AI Institute research
- 2026 priority areas: trustworthy/safe AI, AI for science and discovery, AI-human interaction, quantum-AI intersections, and AI for national security
NSF's Commitment to AI Leadership
NSF has made artificial intelligence a cross-cutting priority across its entire portfolio. This isn't just one directorate's program — AI research is woven into computer science, mathematics, biology, social science, education, and engineering funding. If your research involves machine learning, data science, autonomous systems, natural language processing, computer vision, or the societal impacts of AI, there is an NSF funding pathway available to you. The question isn't whether NSF funds your kind of AI work — it's which program fits your specific research stage and institutional context.
National AI Research Institutes: The Flagship Investment
NSF National AI Research Institutes are the centerpiece of the federal AI research strategy — large, multi-institution collaborative centers funded at $16–20 million each over 4–5 years. As of 2026, NSF has funded over 25 AI Institutes in partnership with other federal agencies (USDA, DOD, DHS, DOE, NIH) covering AI applications in climate science, agriculture, healthcare, manufacturing, transportation, education, and national security. Each Institute involves a lead university and typically 10–30 partner institutions, including community colleges and MSIs.
If your institution is not already an AI Institute participant, the path in is either waiting for a new competition (NSF announces these via Dear Colleague Letters on NSF.gov) or seeking a partnership role with an existing Institute. New Institute competitions are selective — applicants need strong preliminary work, diverse teams, and a compelling case for why the proposed Institute addresses a national AI challenge not covered by existing institutes.
| Program | Award Range | For Whom |
|---|---|---|
| National AI Research Institutes | $16–20M over 4–5 years | University consortia; multi-institution teams |
| CAREER Award (AI-focused) | $500K–$600K over 5 years | Early-career faculty (pre-tenure) |
| CISE Core Research (SHF, CNS, IIS, CCF) | $500K–$1.5M typical | Individual/small team investigators |
| ExpandAI Partnership | Up to $1M over 2 years | Minority-Serving Institutions |
| Convergence Accelerator (AI tracks) | $750K Phase I; up to $5M Phase II | Multi-sector teams with use-inspired focus |
| AI in STEM Education (CS-EDHDI) | Varies | Education researchers + institutions |
| CISE Research Initiation Initiative (CRII) | Up to $175K over 2 years | Early-career researchers at non-R1 institutions |
For Early-Career AI Researchers: CAREER and CRII
The NSF CAREER Award is the most prestigious early-career grant in academic science — $500,000–$600,000 over five years for pre-tenure faculty with outstanding research and teaching plans. AI is one of the highest-funded CAREER categories in the CISE directorate. CAREER applicants must be in tenure-track faculty positions; there are three annual submission windows (July, October, January — varies by directorate).
For early-career researchers at primarily undergraduate or teaching-focused institutions, the CISE Research Initiation Initiative (CRII) provides up to $175,000 over two years to help faculty establish research programs — with no PhD students required. This program explicitly helps researchers at institutions without large PhD programs compete for NSF CISE funding. Many CRII awardees later receive full CAREER awards.
NSF's 2026 AI Priority Research Areas
While NSF funds AI research broadly, program officers and review panels in 2026 have been particularly receptive to proposals in several areas:
Trustworthy and Safe AI — robustness, interpretability, privacy-preserving AI, watermarking and provenance, detection of AI-generated content. This area has explicit congressional and White House interest driving NSF investment.
AI for Scientific Discovery — AI methods applied to fundamental scientific problems: protein structure prediction, drug discovery, materials science, climate modeling, particle physics. NSF cross-directorate programs fund AI/science convergence.
AI-Human Interaction and Societal Impacts — human-AI teaming, AI in high-stakes decision-making (healthcare, criminal justice, hiring), workforce impacts of AI automation. The Directorate for Social, Behavioral and Economic Sciences (SBE) has growing AI program portfolios.
Quantum-AI Intersections — quantum machine learning, quantum advantage for AI tasks, quantum sensing and AI. NSF Quantum Leap Challenge Institutes intersect with AI in several programs.
ExpandAI: Building Capacity at Minority-Serving Institutions
NSF ExpandAI is designed to ensure that the benefits of AI research investment reach Minority-Serving Institutions — HBCUs, HSIs, Tribal Colleges, and other MSIs. ExpandAI partnership grants provide up to $1 million over two years for MSIs to partner with existing NSF AI Research Institutes, participate in collaborative research, and build faculty and student capacity in AI. ExpandAI also funds planning grants for institutions building toward full AI research programs. If you are at an MSI, ExpandAI is often a more accessible entry point to NSF AI funding than competing directly for Institute leadership roles.