Accelerator

AI Safety Grants and Funding Sources: A Complete Directory

Most AI safety founders waste months chasing funders who don't fund their stage, their geography, or their model. This is a complete directory of where to actually look, from the largest philanthropic programs to government grants and venture capital with AI safety mandates.

← Back to AI Safety Startup Guide

AI safety funding has grown substantially over the past five years, but it remains concentrated and competitive. Knowing where to look, what each funder prioritizes, and how to position your work for the specific funder's criteria can dramatically improve your odds. This directory covers the major funding categories and programs as of mid-2026.

Philanthropic grants

Open Philanthropy

Open Philanthropy is the largest single philanthropic funder of AI safety work globally. They fund across technical alignment, interpretability, governance, policy, and field-building. Grant sizes range from individual researcher fellowships (USD 100k-300k per year) to large institutional grants in the millions. They fund both nonprofits and for-profit companies, and they have an informal exploratory grant process with no fixed application deadline. The best entry point is reaching out to their AI safety program officers directly.

Long-Term Future Fund (LTFF)

The LTFF is an EA Funds program that makes smaller, faster grants to individuals and early-stage projects working on reducing existential risk from AI. It is particularly useful for individual researchers transitioning to independent work, early-stage research projects, and small organizations that are too early for Open Philanthropy. Applications are via EA Funds and reviewed on a rolling basis.

Survival and Flourishing Fund (SFF)

SFF makes grants specifically focused on reducing existential risks from AI and other catastrophic risks. They operate through periodic grant rounds (typically twice per year) and are open to both nonprofits and commercial ventures. Their process includes a public application followed by a funding round where regrantors vote on allocation.

Manifund

Manifund is a platform for impact-focused funding that combines philanthropy and impact investing. It enables AI safety researchers and founders to receive funding from multiple sources through a single application. It is particularly useful for smaller amounts and for finding early community validation alongside funding.

Government grants

UK: Innovate UK and ARIA

The UK government's Innovate UK runs regular grant competitions for companies developing AI safety-relevant technologies. The Advanced Research and Invention Agency (ARIA) is a new agency modeled on DARPA that is actively funding early-stage AI safety and trustworthy AI research. Grants are non-dilutive and typically range from GBP 50k to GBP 3M for company projects.

US: NSF, DARPA, and SBIR/STTR

The US National Science Foundation has several programs relevant to AI safety including the Ethical and Responsible AI program and the Trustworthy AI initiative. DARPA funds high-risk technical research including AI interpretability and robustness. The SBIR (Small Business Innovation Research) and STTR (Small Business Technology Transfer) programs provide non-dilutive grants of up to USD 2M in two phases to small companies conducting R&D in areas of national need.

EU: Horizon Europe

EU Horizon Europe runs specific calls for trustworthy AI, AI safety, and AI governance research. Grants range from EUR 500k to EUR 10M and are available to companies, universities, and nonprofits across EU member states and associated countries. Application cycles are typically once or twice per year per topic.

Venture capital

For commercial AI safety ventures, several VC funds are actively looking at this space:

How to position your application

Across every funding category, the most important positioning element is a clear causal theory of impact: how does your work, specifically, reduce the probability of AI harm? The more concrete and falsifiable your theory, the stronger your application. Avoid vague commitments to "beneficial AI" in favor of specific claims: "Our interpretability tool will let safety researchers identify X type of misalignment before deployment, which reduces the probability of Y outcome by making Z monitoring possible."

The Better Societies Accelerator connects founders with the AI safety funding ecosystem. Apply to the accelerator.

Related reading

Frequently asked questions

What is the largest funder of AI safety work?

Open Philanthropy is the largest single philanthropic funder of AI safety work globally, having committed hundreds of millions of dollars to technical research, governance, and field-building across alignment, interpretability, and policy domains.

How do I apply for an AI safety grant?

The process varies by funder. Open Philanthropy has a rolling exploratory process. The LTFF has rolling applications through EA Funds. Government grants have specific calls with fixed deadlines. The most important step for any application is clearly articulating how your work reduces AI risk through a specific, credible causal pathway.

Can for-profit AI safety companies receive grants?

Yes. Open Philanthropy funds both nonprofits and for-profit organizations. Government grants in the UK (Innovate UK), US (NSF SBIR/STTR), and EU (Horizon Europe SME Instrument) are specifically designed for companies. Philanthropic funders are increasingly open to impact investing for safety-focused commercial ventures.

Apply to the Better Societies Accelerator

Built for founders working at the intersection of AI safety and real-world impact. Curriculum, community, and support to help you find the right funding path.