nih-billions-ai-research-how-to-apply

title: 'NIH Is Spending Billions on AI Research — Here's How to Get Your Share' description: 'NIH funded $2.3 billion in AI and ML research in FY2023 alone. Here is how researchers and institutions can find and apply to the growing pool of AI-tagged NIH FOAs.' date: '2026-02-24' author: 'Claire Cummings'

The numbers are hard to argue with. Between fiscal years 2019 and 2023, NIH-funded AI and machine learning research grew 233% in inflation-adjusted dollars — while the overall NIH budget grew just 12%. A cross-sectional study published in JMIR found that by FY2023, the agency was committing $2.3 billion annually to AI and ML projects, covering roughly 3,400 active grants across all 27 institutes and centers. The federal government as a whole now ranks NIH third among all agencies in core AI R&D spending, at $309 million for FY2025.

That trajectory has not reversed. If anything, NIH has redoubled its commitment — launching dedicated programs, approving second stages of flagship initiatives, and distributing AI funding across institutes that span cancer, aging, imaging, cardiology, and behavioral health. The challenge for applicants is no longer whether NIH funds AI research. It is knowing where to look, which mechanisms apply, and how to position work that touches machine learning or computational biology within the right funding channel.

The Flagship Programs: Bridge2AI and AIM-AHEAD

Two NIH-wide programs define the current moment.

Bridge to Artificial Intelligence (Bridge2AI), run through the NIH Common Fund, launched in 2021 with a commitment of $130 million over four years to build ethically sourced, AI-ready biomedical datasets. The premise was simple: AI models in biomedicine are only as good as the training data, and most existing clinical datasets are too small, too narrow, and too poorly documented for robust AI development. Bridge2AI funded four large data-generation centers, a standards and ethics core, and a dissemination and evaluation hub. On January 29, 2026, the NIH Council of Councils approved the program to advance to Stage 2 — a new phase that will move from data creation to application, using the existing datasets to build trusted AI tools for real health challenges. New funding opportunities from Stage 2 have not yet been posted, but researchers should join the Bridge2AI listserv to receive announcements as they open.

AIM-AHEAD (Artificial Intelligence/Machine Learning Consortium to Advance Health Equity and Researcher Diversity) runs a parallel track with a different mandate: ensuring that AI in medicine works for communities historically underrepresented in clinical data. Through its Year 4 programs, AIM-AHEAD is funding Consortium Development projects (FAIR-MED) at up to $800,000 per project over 24 months for community-engaged AI research, and Public-Private Partnership projects capped at $612,500. AIM-AHEAD also runs competitive training programs; the 2025 call for applications targeted institutions serving underrepresented populations. Organizations with strong community health equity track records and EHR data partnerships are well positioned.

Institute-Level AI Funding: Where the Bulk of the Money Lives

The flagship programs are visible, but most NIH AI funding flows through individual institute mechanisms — R01s, R21s, and cooperative agreements where AI or computational methods are embedded in the science rather than treated as the primary subject.

NCI maintains a dedicated AI in Cancer Research funding page that catalogs active opportunities across its divisions. The institute funds AI for imaging analysis, genomic data interpretation, pathology automation, and clinical trial optimization. Most NCI AI work enters through its standard investigator-initiated mechanisms, where reviewers increasingly expect applicants working with large omics or imaging datasets to address computational approaches explicitly.

NIBIB (National Institute of Biomedical Imaging and Bioengineering) runs one of the most active AI grant portfolios at NIH, with funded projects spanning deep learning for medical image reconstruction, convolutional neural networks for diagnostics, and data reduction pipelines. NIBIB's research funding page for AI/ML lists current opportunities and contact program officers who can advise on fit before submission.

NIA has committed more than $40 million through its Artificial Intelligence and Technology Collaboratories (AITC) for Aging Research, a national network anchored at Johns Hopkins, UMass Amherst, and the University of Pennsylvania. The AITC network funds pilot projects up to $200,000 for 12 months, focusing on AI tools for older adults and Alzheimer's disease and related dementias (AD/ADRD). An active funding opportunity for a new AITC center (RFA-AG-26-006, P30 mechanism) is listed on Grants.gov. Investigators who do not hold a P30 center award can still access AITC resources as collaborators.

NIGMS supports computational biology as a core scientific area, and its Biomedical Technology Optimization and Dissemination Program lists computational modeling and AI-adjacent methods among its interests, with receipt dates in January and May.

The full landscape of institute-level AI initiatives — across all 27 ICs — is indexed at the NIH Office of Data Science Strategy. If you do not know which institute funds work in your area, start there.

How to Find Active Opportunities Right Now

NIH changed how it publishes funding opportunities starting FY2026: as of October 2025, Grants.gov is the single official source for NIH FOAs and RFAs. The NIH Guide for Grants and Contracts will no longer post new opportunities. That shift matters for search strategy.

To find active AI-tagged opportunities on Grants.gov, filter by agency (HHS/NIH) and search keywords including "artificial intelligence," "machine learning," "deep learning," or "computational" depending on your domain. Pair that with a search on NIH RePORTER using the "Machine Learning and Artificial Intelligence" funding category — that filter returns funded grants, which tells you which institutes are actively investing and which program officers are managing those portfolios. A cold email to the program officer listed on a funded project in your area is often the most efficient step a researcher can take before drafting an application.

Two application mechanics are worth knowing. First, NIH's new simplified review framework applies to applications due on or after January 25, 2025 — reviewers evaluate Significance, Rigor and Feasibility, and Team and Resources rather than the older five-criterion model. AI-specific applications benefit from explicit discussion of dataset quality, model validation, and responsible AI practices in the Rigor section. Second, a policy notice issued in July 2025 (NOT-OD-25-132) establishes that applications substantially drafted by AI will not be considered original work and limits each PI to six new applications per calendar year — worth factoring into your submission calendar.

Positioning Your Work for Review

The JMIR study offers a useful calibration: 70% of NIH AI/ML grant recipients hold only a PhD, and that group accounts for roughly two-thirds of all AI/ML funding dollars. Biomedical informatics, computational biology, and data science researchers apply and win in roughly equal measure across R01, R21, and U-series cooperative agreements. Interdisciplinary teams — combining domain scientists with AI methodologists — consistently fare well, particularly when the dataset is novel or the clinical translation pathway is clear.

If your research touches patient-level data, reviewers will expect a data governance plan, a discussion of potential biases in training datasets, and some engagement with how the model performs across demographic subgroups. Both Bridge2AI and AIM-AHEAD have driven those expectations into the broader reviewer culture across NIH.

NIH's $2.3 billion AI investment is real, dispersed, and growing — and most of it is accessible through the same R01 and R21 mechanisms that fund any NIH science. Tools like Granted can help you identify the right FOA, match it to your team's profile, and move from a rough specific aims page to a submission-ready application before the deadline.

Sources:

• NIH-Funded AI and ML Research, 2019-2023: Cross-Sectional Study (JMIR)
• Federal AI and IT R&D Spending Analysis (Federal Budget IQ)
• Bridge2AI Program | NIH Common Fund
• Bridge2AI Advances to Next Stage | NIH Common Fund
• AIM-AHEAD Program Overview
• AIM-AHEAD Year 4 Call for Proposals
• AI in Cancer Research Funding Opportunities (NCI)
• AI/ML and Deep Learning | NIBIB
• AITC for Aging Research | NIA
• RFA-AG-26-006 on Grants.gov
• Artificial Intelligence Initiatives | NIH Data Science
• Institute and Center AI Initiatives | NIH Data Science
• NIH RePORTER
• NOT-OD-25-132: Supporting Fairness and Originality in NIH Research Applications