Top 20 Foundations Funding AI Research in 2026 (Ranked by Giving)

Last verified: April 11, 2026.

Private foundations now rival federal agencies as drivers of artificial intelligence research. While NSF invests over $700 million annually in AI, a parallel universe of philanthropic capital -- from Open Philanthropy's alignment labs to the McGovern Foundation's public-purpose deployments -- is reshaping which questions get asked and who gets to ask them.

Browse our AI Grants page for current opportunities, or search 133,000+ foundations on our Foundations directory.

The Billion-Dollar Coalitions

The single largest commitment came in October 2025, when ten foundations launched Humanity AI, a $500 million, five-year initiative co-chaired by the MacArthur Foundation and Omidyar Network. Founding members include the Ford Foundation, Mellon Foundation, Mozilla Foundation, Kapor Foundation, Doris Duke Foundation, Lumina Foundation, Siegel Family Endowment, and the David and Lucile Packard Foundation. Pooled-fund grants begin in 2026, targeting democracy, education, labor, cultural preservation, and national security. Rockefeller Philanthropy Advisors serves as fiscal sponsor.

Separately, the Patrick J. McGovern Foundation has emerged as the single most concentrated AI funder in philanthropy. In 2025 alone, it committed $75.8 million across 149 grants to organizations deploying AI for climate resilience, health equity, crisis response, and digital literacy in 13 countries. Total giving over the past decade exceeds $500 million, placing it among the largest supporters of public-purpose AI anywhere.

AI Safety and Alignment

Open Philanthropy stands alone as the dominant private funder of AI safety research. Since 2017, it has directed roughly $336 million toward reducing risks from advanced AI -- approximately 12% of its total $2.8 billion in grants. In 2025, it issued a landmark request for proposals committing $40 million over five months to technical AI safety research across 21 areas, from interpretability to dangerous-capability evaluations. Annual AI safety spending has hovered near $46-50 million, making it the largest single funder in the field by a wide margin.

Schmidt Sciences, the philanthropic organization founded by Eric and Wendy Schmidt, runs three AI-specific programs. Its AI2050 Fellows program awarded $18 million to 28 scholars in 2025. The Humanities and AI Virtual Institute (HAVI) distributed $11 million to 23 research teams applying AI to archaeology, history, and literature, with a new round accepting proposals through March 2026 for grants up to $800,000. A Science of Trustworthy AI RFP, due May 17, 2026, funds technical research on understanding and controlling frontier AI risks.

The Mozilla Foundation channels funding through its Democracy x AI cohort, fellowships, and an experimental incubator. The 2026 Fellows Program supports up to 10 leaders with $100,000 each ($75,000 stipend plus $25,000 project budget). Its $1 million AI & Democracy awards will fund 10 projects over 24 months focused on information ecosystem resilience and algorithmic accountability.

Science, Math, and Fundamental Research

The Simons Foundation, with assets exceeding $5 billion and annual grants of approximately $180 million, funds AI at the intersection of mathematics and computation. Its Targeted Grants in Mathematics and Physical Sciences support high-risk theoretical projects for up to five years on a rolling basis. Alongside NSF, Simons co-funded a $6.6 million Institute for Computer-Aided Reasoning in Mathematics at Carnegie Mellon, and its SIAM-Simons Undergraduate Research Program accepts applications for summer 2026.

The Gordon and Betty Moore Foundation distributed $401.8 million across 952 grants in 2024, with individual awards ranging from $50,000 to $5 million. Its Moore Inventor Fellows program -- $200,000 per year for three years, with nearly $34 million committed through 2026 for 50 fellows -- increasingly supports researchers at the AI-hardware boundary. A recent $250,000 grant to the ARISE research network funds clinical trials evaluating AI-clinician collaboration in complex medical decisions.

The Alfred P. Sloan Foundation disburses roughly $80 million annually across science, technology, and economics. Its Metascience and AI Postdoctoral Fellowship offers up to $250,000 over two years for early-career researchers studying how AI transforms the research ecosystem, with a residential summer school planned for 2026. Sloan's Exploratory Grantmaking in Technology program accepts letters of inquiry on a rolling basis.

The Kavli Foundation provides approximately $35 million in annual grants across astrophysics, nanoscience, neuroscience, and science-and-society. Its 2026 AI-for-Science Fellowship, run through FutureHouse, supports independent postdoctoral research in AI-enabled neuroscience. The NeuroData Discovery Grants fund creative reanalysis of open neuroscience datasets. Kavli does not accept unsolicited proposals but issues periodic open calls.

AI for Society and Governance

The Bill & Melinda Gates Foundation -- $8.6 billion in annual spending, scaling to $9 billion by 2026 -- funds AI through its Grand Challenges program and a dedicated AI Fellows Program focused on global health, development, and education. More than 50 grants currently support AI innovators in low- and middle-income countries tackling health challenges.

The Ford Foundation committed to the $200 million philanthropic coalition for public-interest AI alongside Omidyar, Hewlett, and others, and provides multiyear general operating support to AI governance organizations including the AI Now Institute and Partnership on AI.

The William and Flora Hewlett Foundation funds AI through large institutional grants: $5 million to FAMU's Cyber Policy Institute, $5 million to Georgetown's Center for Security and Emerging Technology, $2.5 million to Vanderbilt's National Security Institute for AI-driven information warfare research, and $1 million to the Data-Driven EnviroLab for AI climate work.

Omidyar Network launched a dedicated $30 million investment in responsible generative AI in March 2025, targeting inclusive AI infrastructure, diverse voices in AI governance, and competition in AI markets. Its Responsible Technology Youth Power Fund distributed $2.4 million to 18 youth-led organizations, and a Tech Journalism Fund offers $5,000-$25,000 grants for in-depth AI reporting.

The Rockefeller Foundation has invested more than $18 million in public-sector AI, including the AI Readiness Project that aims to expand from 30 to all 50 states by 2026. It supports at least ten state-level pilots testing responsible AI in government services.

The John S. and James L. Knight Foundation funds AI at the journalism-technology intersection: $4 million (with USC Marshall match) for purpose-driven AI research, $2 million for a San Jose AI Center of Excellence through Plug and Play, $1 million to Northwestern for generative AI in newsrooms, and up to $20,000 research grants through the Authors Alliance for AI-copyright scholarship.

How to Win a Foundation AI Grant in 2026: The 5-Step Playbook

Most foundation AI funding clusters around four themes: safety and alignment (Open Philanthropy, Schmidt, Mozilla), fundamental research (Simons, Moore, Sloan, Kavli), public-purpose deployment (McGovern, Gates, Rockefeller), and governance and rights (Ford, Hewlett, Omidyar, Knight). Identifying which cluster a foundation belongs to is the easy part. Actually winning the grant requires a different playbook than the one most researchers use for federal submissions.

1. Read the foundation's theory of change before writing a single sentence. Every major AI funder has a public theory of change — the causal model connecting what they fund to the outcomes they want. Open Philanthropy wants to reduce catastrophic AI risk by a measurable percentage. The McGovern Foundation wants to deploy AI to underserved populations in 13 specific countries. Gates wants health outcomes in low- and middle-income economies. Simons wants mathematical rigor on problems computer scientists think are impossible. Your proposal has to fit the theory of change explicitly — and cite the funder's own framing language back at them.

2. Pick the right mechanism. Most foundations have three or four different ways to give you money. Open Philanthropy has open Requests for Proposals, grant opportunities listed on their site, and relationships-based giving for repeat grantees. Schmidt Sciences has AI2050 (for senior fellows), HAVI (for humanities-AI research), and the Science of Trustworthy AI RFP (for technical safety). If you apply to the wrong mechanism, you will get a polite no even when your work is a perfect fit for a different program at the same foundation. Read every program description before you submit.

3. Lead with a letter of inquiry, not a full proposal. The foundations above with rolling LoI policies — Simons, Sloan, Kavli, Hewlett, Ford — are telling you something: they want to read a 2-page pitch before you invest a month writing a full proposal. Use the LoI as an audit. If you cannot get a positive response to a strong LoI, either your framing is off or the program is not for you. Save the full proposal effort for the programs that respond with "send us more."

4. Show traction, not promise. Foundations fund credibility at least as much as they fund ideas. A researcher with one AI-safety paper and a credible collaborator will beat a researcher with three papers and no collaborators, all else equal. Before you submit, collect: the two or three most relevant publications, one existing funder or institutional backer who will vouch for you, and one named collaborator from the foundation's past grantee list. Foundations check grantee alumni networks before they write checks.

5. Match the deadline calendar, not the fiscal-year calendar. Foundation deadlines are not synced to your institution's fiscal year or NIH's study-section cycle. Open Philanthropy runs open windows of weeks or months. Schmidt Sciences' Trustworthy AI RFP closes May 17, 2026. Moore's Inventor Fellows typically open April through July. Sloan and Simons accept rolling letters of inquiry year-round. The researchers who win foundation grants consistently maintain a rolling deadline calendar for the 15-20 foundations most likely to fund their work, and they start each proposal 8-12 weeks before the deadline — not two.

Where Granted fits

Mapping 133,000 foundations against your specific research profile by hand is the part of the process that kills proposals before they start. Granted indexes every foundation named above from IRS 990 filings — plus 132,985 others — with giving patterns, grant size distributions, board members, and historical recipients. Search by keyword, by geographic focus, or by average grant size, and export a prospect list of the 20 foundations most likely to fund your work in minutes. From there, you write the LoI, not the prospect list.

For researchers and nonprofits navigating the foundation AI landscape, Granted can surface relevant foundation opportunities and help you move from a concept note to a polished submission before the window closes.