Open Philanthropy is a San Francisco-based philanthropic grantmaker funded principally by Cari Tuna and Dustin Moskovitz. It identifies and funds cause areas it judges to be neglected, tractable, and high-impact, and across the 2015–2026 period it has been the most consistent institutional funder of the AI safety and AI governance research ecosystem. On November 18, 2025 the organization rebranded as Coefficient Giving, keeping its mission and grantmaking approach while signaling a shift from a single-anchor-donor model toward working with multiple donors (Source: coefficientgiving.org). Because most of the field still refers to it as Open Philanthropy, both names are in use; this page uses the historical name except where the rebrand is the point.
| Field | Value |
|---|---|
| Type | Philanthropic grantmaker (LLC with 501(c)(3) affiliate) |
| Headquarters | San Francisco, California |
| Founded | 2014 (as a joint project of GiveWell and Good Ventures; independent since 2017) |
| Principal funders | Cari Tuna and Dustin Moskovitz (Facebook co-founder; Asana) |
| Longtime lead | Holden Karnofsky (co-founder; later at the Carnegie Endowment and Anthropic — see below) |
| Known for | Being the largest institutional funder of the modern AI safety and AI governance research ecosystem |
Overview
Open Philanthropy identifies and funds cause areas it characterizes as neglected, tractable, and high-impact. Its Global Catastrophic Risks portfolio includes biosecurity and AI safety and governance. Its longtermism framing, drawn from the effective altruism tradition, has shaped the field's self-conception since the late 2010s.
Activities
Across the 2015–2026 period, Open Philanthropy has been the most consistent institutional funder of several parts of the AI safety field:
- Technical AI safety research, through grants to ARC, Redwood Research, Apollo Research, METR, CAIS, and numerous academic groups.
- Early-stage AI companies focused on safety, most prominently a 2017 grant to OpenAI (with an attendant board seat) and later significant investment in Anthropic.
- AI policy research institutions, including a founding grant to CSET, plus GovAI, Epoch AI, and the broader governance field.
- Talent pipeline programs, including Horizon Institute for Public Service fellowships and related programs.
- Individual researcher grants to independent researchers and early-career PhDs.
Most independent US and UK AI safety nonprofits have Open Philanthropy in their funder lists. The organization does not typically produce research itself; its role is financial. Its public-facing writing, such as grant writeups and Karnofsky's blog posts, is nonetheless influential in the field, and the organization functions as financial infrastructure for the field rather than as a research producer.
People
- Holden Karnofsky — co-founder; longtime lead of the AI worldview and strategy work; author of the May 2016 post that introduced the term "transformative AI" (Some Background on Our Views Regarding Advanced Artificial Intelligence (Karnofsky, 2016)) and of the "Most Important Century" blog series framing the stakes of transformative AI. In 2024 he moved to work on AI safety directly, in roles adjacent to Anthropic; the RE-Bench paper of November 2024 lists him with an Anthropic affiliation and notes that his work on it was done while at the Carnegie Endowment for International Peace (RE-Bench: Evaluating frontier AI R&D capabilities of language model agents against human experts (Wijk et al., 2024)).
- Alexander Berger — CEO since 2021.
- Luke Muehlhauser — longtime lead of the AI governance portfolio.
Grantmakers on AI safety rotate. The staff overlap heavily with the effective altruism and rationalist-adjacent professional networks.
Positions
Open Philanthropy announced in 2016 that it planned to make potential risks from advanced AI "a major priority" that year. Karnofsky's background post for that decision introduced the term "transformative AI," defined as AI that "precipitates a transition comparable to (or more significant than) the agricultural or industrial revolution," and put the probability of such AI within 20 years at at least 10%. He wrote that the cause would probably have remained a focus area without that estimate, but that the estimate was important to prioritizing it as highly as planned, and that the post's reasoning drew on the organization's "philosophy of hits-based giving" (Some Background on Our Views Regarding Advanced Artificial Intelligence (Karnofsky, 2016); see Transformative AI (TAI)).
Open Philanthropy's AI work treats catastrophic and existential risk as first-order concerns, more so than mainstream philanthropy, and this framing extends through the ecosystem it funds. Unlike purely advocacy-focused funders, it has funded both empirical alignment research and policy institutions.
The organization's role has drawn debate. Critics argue that its dominance has homogenized the AI safety field, concentrated epistemic authority within the EA-adjacent community, and made the field's research agenda legible mainly to its donors. Defenders argue that the field existed only because Open Philanthropy built it.
After the 2022 FTX collapse, which had co-funded parts of the AI safety ecosystem, Open Philanthropy took on increased responsibility and tightened some grantmaking practices.
Funding and governance
The principal backers are Cari Tuna and Dustin Moskovitz. Moskovitz co-founded Facebook and Asana and holds most of the couple's wealth in Asana and other tech holdings. Open Philanthropy functions as the policy- and science-focused arm of the Tuna and Moskovitz philanthropic project; a parallel organization, Good Ventures, is the grantmaking vehicle.
AI safety funding amounts are large and have grown substantially since 2022, though public figures on specific portfolios are partial.
In June 2025, Good Ventures announced it would stop supporting certain sub-causes and would not expand into new cause areas by default, a change that caused some grantees to lose funding (Source: forum.effectivealtruism.org). The November 2025 rebrand to Coefficient Giving was framed as a move beyond the single-anchor-donor structure — historically dominated by Tuna and Moskovitz's wealth — toward a model that pools and advises capital from multiple donors, while keeping AI safety, biosecurity, and the other global-catastrophic-risk portfolios (Source: forbes.com). The shift coincides with an anticipated wave of philanthropic capital from Anthropic employees pledging to give away equity gains, much of which is expected to flow toward the effective-altruism-aligned AI-safety nonprofits the organization helped build (Source: transformernews.ai).
Relationships
- related: Alignment Research Center (ARC), Redwood Research, Apollo Research, METR, Center for AI Safety (CAIS) — funded organizations
- related: Center for Security and Emerging Technology (CSET), Centre for the Governance of AI (GovAI), Epoch AI — funded policy research institutions
- related: Horizon Institute for Public Service — talent pipeline funding partner
- related: Anthropic, OpenAI — equity and grant relationships, historically consequential
- related: Future of Life Institute (FLI), Machine Intelligence Research Institute (MIRI) — adjacent but differently-funded existential-risk organizations
- related: Holden Karnofsky — co-founder; Some Background on Our Views Regarding Advanced Artificial Intelligence (Karnofsky, 2016) — the 2016 post introducing "transformative AI"
- supports: Transformative AI (TAI) — originator of the term
- supports: AI Safety Cases and Frameworks, Compute Governance, AI Scheming — fields Open Philanthropy's funding built out