AI Grants for Nonprofits: A Complete Guide

Somewhere between the TED talks about AI transforming every industry and the reality of a nonprofit running donor management on a ten-year-old spreadsheet, there is a funding gap. It is enormous, and it is widening. The social sector is being told to adopt AI or risk irrelevance, but the organizations doing the telling rarely mention who is supposed to pay for it.

The good news: a growing number of funders have noticed the disconnect. Federal agencies, tech company foundations, and private philanthropies are now directing real money toward nonprofit AI adoption -- not just the kind that funds academic research labs, but the kind that helps a food bank automate intake screening or a legal aid nonprofit triage case referrals. Browse current AI funding opportunities on our AI Grants page.

Here is where to find it.

The Foundation Landscape: Who Is Actually Writing Checks

The largest and most accessible funding source for nonprofit AI work is not the federal government. It is a handful of foundations that have made AI adoption in the social sector a core priority.

The Patrick J. McGovern Foundation committed $75.8 million across 149 grants in 2025 alone, funding organizations applying AI to climate resilience, human rights, health equity, and crisis response across 13 countries. Individual awards have ranged from $500,000 to $750,000, with recipients including the ACLU Foundation and Amnesty International. What distinguishes McGovern from most tech-oriented funders is that it pairs financial support with in-house technical teams who advise grantees on data governance, model evaluation, and risk assessment -- the exact kind of capacity a nonprofit running its first AI project needs.

Google.org runs a Generative AI Accelerator that provides 20 participants a share of $30 million alongside six months of pro bono support from Google engineers, technical training, and Google Cloud credits. The 2025 cohort tackled crisis response, antimicrobial resistance, and children's mental health. Google.org has also deployed $75 million through its AI Opportunity Fund, focused on AI workforce development and education nonprofits, including $10 million specifically earmarked for technical assistance to help nonprofits adopt AI tools at no cost.

The KPMG U.S. Foundation is distributing $6 million in AI integration grants to U.S. nonprofits, coupled with AI learning sessions, pro bono consulting, and a collaboration with Salesforce to build responsible AI capacity across the sector. Salesforce itself has committed $17 million to 18 nonprofits for AI readiness, plus a separate Accelerator providing $2 million in funding with 24-month donated Salesforce product access to help six organizations build AI-driven solutions.

The OpenAI Foundation's People-First AI Fund disbursed $40.5 million in unrestricted grants to 208 U.S.-based nonprofits in 2025, with an additional $9.5 million in board-directed grants. Eligibility required 501(c)(3) status and annual operating budgets between $500,000 and $10 million -- a sweet spot that captures many mid-sized nonprofits typically overlooked by both small community foundations and large federal programs. The fund explicitly welcomed organizations at every stage of AI adoption, from exploration through active deployment.

Federal Programs Worth Watching

Federal AI funding skews heavily toward research institutions and defense contractors, but several programs create openings for nonprofits.

The NSF Civic Innovation Challenge (CIVIC) funds partnerships between researchers and community organizations to deploy emerging technologies, including AI. Stage 1 planning grants run up to $75,000, with Stage 2 full awards reaching $1,000,000. The program specifically targets projects that transition foundational research into community-level practice -- exactly the kind of applied AI work most nonprofits need funded.

NSF's broader AI portfolio, including the $100 million expansion of its National AI Research Institutes program, occasionally includes pathways for nonprofits partnering with universities. The key is positioning your organization as the community implementation partner in a joint proposal rather than competing as a standalone applicant against research universities.

NIST's Small Business Innovation Research (SBIR) program has allocated over $3 million to AI-focused projects, though these target small businesses rather than traditional nonprofits. Social enterprises with a for-profit arm may find a fit here, particularly in Phase I awards that fund six-month feasibility studies with up to $400,000 available in Phase II follow-on funding.

Closing the Gap Between Hype and a $30K Pilot

Most nonprofits do not need $5 million for a research institute. They need $20,000 to $50,000 for a pilot: a proof of concept that automates one painful workflow, improves one intake process, or surfaces one pattern in program data that no one had time to look for manually.

This is where capacity-building and technical assistance programs matter more than large competitive grants. Google.org's $10 million investment in Project Evident and the Tech:NYC Foundation funds in-person working sessions, webinars, and peer learning for nonprofits exploring AI implementation -- at no cost to participants. The KPMG Foundation's AI 101 sessions have already trained over 180 nonprofit leaders from 80 organizations on practical AI applications.

For organizations with annual budgets under $500,000 -- too small for the OpenAI Fund, too mission-driven for SBIR -- community foundations and regional grantmakers are increasingly adding AI capacity-building to their portfolios. Fidelity Charitable has published guidance encouraging donors to direct philanthropic funds toward nonprofit technology adoption. State-level programs, like Maryland's Cyber and AI Pilot Clinic Grant Initiative offering up to $500,000 per grantee, are emerging as well.

The practical move for a nonprofit considering its first AI project is not to wait for the perfect grant. It is to start with a clearly defined problem, estimate the real cost of a pilot, and then match that number against the programs listed above. A $30,000 planning grant from a community foundation or a free technical assistance program from Google.org can get a proof of concept running. That proof of concept becomes the strongest evidence in a larger application to McGovern, OpenAI, or a federal program down the road.

What Funders Want to See in an AI Proposal

Across all of these programs, certain elements consistently surface in successful applications. Funders want to see a specific, bounded problem -- not "we want to use AI" but "we want to reduce case intake processing time by 40% using document classification." They want evidence that your organization has the data infrastructure, however basic, to support the proposed work. They want a plan for what happens when the grant ends.

Responsible AI practices have also become a near-universal requirement. The McGovern Foundation provides grantees with risk assessment frameworks. Google.org's accelerator includes responsible AI training. Even smaller funders increasingly ask applicants to address bias, privacy, and transparency in their proposals. If your application does not address these concerns, it is incomplete.

Finally, every funder in this space is looking for scalability and replicability. A pilot that solves one organization's problem is interesting. A pilot that produces a model or toolkit other nonprofits can adapt is fundable.

For nonprofits ready to move from the AI conversation to an actual funded project, Granted can help match your organization's work to the right opportunities and turn a rough concept into a submission-ready proposal.