We Analyzed 4,000+ AI Grants — Here's What We Found

Federal AI funding has crossed a threshold that most grant seekers haven't fully registered yet. The U.S. government's requested investment in AI research and development hit $3.3 billion in FY 2025, up from $3.1 billion the year before — and that figure only counts the NITRD-tracked R&D budget, not the broader wave of agency-specific programs layered on top. We decided to look at what that spending wave actually produces in terms of fundable opportunities.

Granted's database now indexes over 87,000 grants from 144 sources. When we filtered for artificial intelligence, machine learning, and adjacent keywords, we pulled 4,127 distinct opportunities. Browse our Technology & Innovation Grants page for current AI-related opportunities.

Here is what the data told us.

NSF Dominates the Count, but DOE Moves the Biggest Checks

The National Science Foundation leads all agencies in sheer volume of AI-tagged awards, with over $700 million invested annually across its AI portfolio. NSF's centerpiece is the National AI Research Institutes program, which now funds 25 institutes and received a $100 million expansion in July 2025 covering five new institutes focused on materials discovery, drug development, mental health, human-AI collaboration, and STEM education. Individual institute awards run $4 million to $20 million over five years — significant but not economy-reshaping.

The Department of Energy has taken a different approach: fewer awards, far larger envelopes. DOE's Genesis Mission, announced in November 2025, directed $320 million in initial awards across four workstreams — an American Science Cloud for shared compute infrastructure, a Transformational AI Models Consortium led by Argonne National Laboratory, 37 foundational AI awards for domain-specific scientific models, and 14 robotics and autonomous laboratory projects. Congress separately appropriated $150 million through September 2026 for transformational AI models flowing through the same channels.

DARPA allocated $314 million to AI R&D in FY 2025, spread across programs like AI Forward for trustworthy national-security AI and AI FORGE for adapting frontier models to defense applications. NIH committed $309 million, much of it channeled through initiatives like Bridge2AI and AIM-AHEAD.

In our database, the top five agencies by AI grant count are NSF, DOE, NIH, DARPA, and USDA — in that order.

Award Sizes Vary Wildly by Program Type

The average AI grant does not exist in any meaningful sense. A Phase I SBIR award for an AI startup caps at roughly $300,000 through NSF and $314,363 through most other agencies. A Phase II can reach $1.25 million at NSF and over $2 million elsewhere. By contrast, DOE's Genesis Mission consortium awards start at $30 million, and the full Office of Science solicitation (DE-FOA-0003600) covers $500 million in annual awards across all program areas.

NSF's Smart Health and Biomedical Research in the Era of AI program (NSF 25-542) allocates $15 million to $20 million per cycle for AI-health crossover research. Individual investigator awards in this program typically range from $300,000 to $1.2 million over three to four years.

What the data reveals is a bimodal distribution: a large cluster of investigator-scale awards between $150,000 and $500,000, and a smaller but growing cluster of infrastructure and consortium awards above $10 million. The middle ground — multi-PI collaborative projects in the $1 million to $5 million range — exists but represents a smaller share of total opportunities than you might expect.

The Five Fastest-Growing AI Verticals

Looking at year-over-year growth in our database, five application domains are adding AI-tagged grants faster than the overall average.

Healthcare and biomedical AI leads the pack. NIH's 27 institutes are each running their own AI/ML initiatives, and the combined pipeline now stretches across drug discovery, clinical imaging, electronic health records, and health equity. The AIM-AHEAD consortium alone is building AI capacity at minority-serving institutions nationwide.

Energy and climate applications rank second, driven almost entirely by DOE. The Genesis Mission's 26 priority challenges span fusion energy, critical materials, advanced manufacturing, and decarbonization — all flagged for AI integration. DOE separately committed $29 million for AI in fusion energy sciences and $30 million for AI-accelerated grid interconnection.

Defense and national security holds third position by funding volume, anchored by DARPA's AI Forward and the broader Pentagon AI budget that extends well beyond DARPA's $314 million into service-specific research arms.

Education and AI literacy is emerging as a distinct funding vertical. NSF's July 2025 institute cohort included the Institute for Student AI-Teaming at the University of Colorado Boulder, dedicated to AI-supported group learning and semester-long AI literacy curricula.

Materials science and advanced manufacturing rounds out the top five. Cornell's new NSF AI-Materials Institute is tasked with accelerating discovery of next-generation materials for energy, sustainability, and quantum technologies — a category that barely existed in federal AI funding five years ago.

What Changed in 2025 — and What It Means for 2026

Three structural shifts stand out. First, AI funding is no longer concentrated in a handful of programs at NSF and DARPA. Every major science agency now runs dedicated AI initiatives, and the cross-agency coordination through NITRD means applicants can find AI-relevant solicitations at DOE, NIH, USDA, NIST, and even the Department of Education.

Second, the scale of individual awards is climbing. The Genesis Mission's $320 million initial tranche, NSF's $100 million institute expansion, and Congress's $150 million AI models appropriation all represent bets on large, multi-institutional projects. Solo investigators can still compete — but increasingly through programs that fold them into larger collaborative structures.

Third, the SBIR/STTR pipeline for AI startups faces genuine uncertainty. Congressional authorization for SBIR and STTR expired on September 30, 2025, and as of February 2026, reauthorization remains incomplete. Agencies may delay or cancel AI-focused small business solicitations until Congress acts. Startups should monitor sbir.gov and consider DOE and NSF direct-to-investigator programs as interim alternatives.

For researchers and organizations positioning proposals in 2026, the data points to a clear strategy: identify which agency's AI vertical aligns with your domain expertise, target the program-specific solicitations rather than general funding announcements, and build collaborative teams that match the scale funders are increasingly rewarding. Platforms like Granted can surface those matched opportunities and help you move from a database search to a submission-ready proposal.