Coefficient Giving boosts pledge to $1 billion, banking on an AI windfall

Coefficient Giving is writing bigger checks. Last week, the grant maker and philanthropy adviser alongside GiveWell, a charity evaluator and funder, announced it would make a $276 million grant to fund the purchase of roughly 90 million insecticide-treated bed nets for the Democratic Republic of Congo, an effort GiveWell estimates could prevent about 67,000 deaths, most of them of children under 5. 

It’s the first major grant funded as part of its recent decision to raise its 2026 commitment to GiveWell’s recommended global health groups from $175 million to $1 billion. That’s a nearly sixfold jump for the San Francisco organization, which was formerly known as Open Philanthropy and is funded by Good Ventures, the foundation of Facebook co-founder Dustin Moskovitz and Cari Tuna. The group has directed a total of about $7 billion since 2014. 

Coefficient Giving has deep roots with GiveWell. GiveWell incubated what eventually became Coefficient Giving starting in 2011 before it spun off as an independent organization in 2017. They’re separate today but still closely tied. Tuna sits on GiveWell’s board, and GiveWell CEO Elie Hassenfeld sits on Coefficient Giving’s board.

That closeness traces back to the shared philosophical starting point of effective altruism, a movement built on using evidence and cost-effectiveness analysis to determine how to have the most impact with a given dollar. Coefficient Giving still evaluates causes largely on effective-altruism-style criteria, such as how many people are affected, how neglected the cause is, and how difficult or easy it is to make progress. 

Coefficient Giving’s leaders say the timing of this latest funding surge anticipates a possible windfall from the massive boom in artificial intelligence. If OpenAI, Anthropic, and others go public, this funding approach could lay the groundwork for more tech industry donors to “emerge and pick up the baton,” the group wrote in a recent blog post. Grantees will need time to make more ambitious plans, hire new staff, and expand into new regions, they said. 

“Coefficient Giving can smooth the pathway for nonprofits to build capacity in advance,” said Mike Levine, a spokesman. If AI assets become liquid within the next year, giving could jump sharply, and Coefficient Giving wants that transition to be smooth rather than sudden, he said. 

Just how large that eventual wave of wealth could be and what it might mean for philanthropy is now the subject of debate across the nonprofit world. 

One widely discussed and disputed estimate comes from Nan Ransohoff, head of public goods at fintech company Stripe. If OpenAI and Anthropic go public, she wrote in a Substack post, it could unlock $37 billion to $100 billion in new annual philanthropic spending from OpenAI Foundation’s 26% stake in OpenAI as well as Anthropic and OpenAI founders who’ve pledged to give away most of their wealth and employees giving through donor-advised funds.

AI philanthropy deserves scrutiny

Coefficient Giving itself is built for exactly this kind of moment. It’s a pooled philanthropy vehicle that is popular in the tech sector. It also has another edge when it comes to attracting AI funders: Anthropic president Daniela Amodei, sister of company founder Dario Amodei, is married to Holden Karnofsky, who co-founded both GiveWell and Open Philanthropy. Karnofsky joined Anthropic last year. 

Given that, Coefficient Giving’s pledge reflects some key interests of those with concentrated tech wealth that could reshape an entire funding category like global health and development almost overnight. That kind of consolidated decision-making power among the tech wealthy deserves scrutiny in its own right, said Hala Hanna, executive director of MIT Solve, an MIT initiative that funds and supports early-stage entrepreneurs using technology to close equity gaps globally. 

It’s easy to lose sight of just how much money is involved and who is deciding where it goes, she said. If you spend $1 a second, it takes 11 days to spend a million and 31 years to spend a billion, and some of these AI entrepreneurs are going to have billions upon billions of dollars, she said. And because that money will be directed by a small number of unelected individuals rather than through public democratic processes, “I do think we deserve a lot of scrutiny here,” she said. “This is big.”

AI-driven economic optimism is an explicit part of Coefficient Giving’s financial logic for moving now rather than later. 

“We and Good Ventures generally expect that transformative AI, if navigated safely, could make the world as a whole massively richer and reduce the need for global health and development in the future,” the organization wrote. 

Is the money even real? 

Not everyone thinks it’s useful to speculate about where the money will ultimately land. David Goldberg, founder and president of Founders Pledge, a group that advises tech entrepreneurs on their philanthropy, cautioned against assuming a single set of priorities will emerge just because a wave of donors share a similar background. 

“If you’ve met 100 tech donors, you’ve met 100 tech donors, and I’ve met way more than that in my time at Founders Pledge,” he said. 

Tech donors share a similar rigor but not the same values, he added, and many will need support figuring out the best ways to give if the projected sums materialize. 

Nicole Taylor, CEO of Silicon Valley Community Foundation, also adds a dash of skepticism to the anticipation brewing around a potential wave of AI philanthropy. These companies haven’t gone public yet, so this AI wealth boom is “all on paper right now,” she said. And the vast majority of AI company staff, who are mostly in their mid-20s to early 30s, aren’t thinking about philanthropy because they’re still heads-down focused on building their careers, Taylor said. 

That pattern predates AI. Even Mark Zuckerberg and Priscilla Chan, who pledged to give away the bulk of their fortune, have donated roughly 6% of their current net worth over the past 15 years, well under half a percent a year, according to a 2025 analysis by the Institute for Policy Studies. Historically, even the most committed billionaire philanthropists have given away far less of their wealth each year than optimistic AI giving projections assume. 

Coefficient Giving’s own money, by contrast, isn’t hypothetical. It’s built on tech wealth that’s already been realized, not stock still waiting to vest or an IPO that hasn’t happened. Moskovitz is a co-founder and the current board chair of Asana, a work management platform, and leads the company’s AI strategy development. The organization acknowledges that its decision to increase giving is tied to the “valuations of some extremely volatile assets,” though the pledge size doesn’t hinge on any particular company’s IPO timeline, Levine said. 

The volatility of the assets is the reason this is a one-time surge, he explained. Even if AI wealth doesn’t ultimately materialize the way Coefficient Giving anticipates, Levine added, GiveWell’s recommended charities “will still have saved and improved many lives in the interim.”

Are nonprofits prepared for AI wealth?

Hanna sees a deeper problem shared by both these recent commitments and effective altruism more generally: They concentrate on causes that are easiest to measure. Coefficient Giving’s global health giving and OpenAI Foundation’s recent $100 million pledge for AI-assisted medical care both focus heavily on infectious disease, an area effective-altruism-aligned funders favor in part because outcomes like lives saved are countable. 

But noncommunicable diseases, not infectious diseases, are the leading cause of death in much of the world and attract far less funding, Hanna said. She calls this “the lamppost effect,” after the old joke about someone searching for their lost keys under a streetlight because that’s where the light is, not where the keys actually fell. 

“You’ve got to weigh impact and effectiveness rigorously while also widening what counts as impact,” said Hanna, whose organization advocates for the kinds of ventures she says get passed over. 

Some of those who anticipate a flood of philanthropic dollars from AI see a very different issue on the horizon. Ransohoff in her Substack asks whether the nonprofit sector can even absorb this much money. The field needs hundreds or thousands of new “philanthropic startups” and grant-making staff to process an estimated 50,000 new grants a year, Ransohoff argues. She did not respond to a request for comment.

The Silicon Valley Community Foundation’s Taylor pushes back on the idea that the nonprofit sector’s existing infrastructure can’t absorb incoming wealth. Waiting for young tech workers to leave full-time jobs and build brand-new giving vehicles wastes time, she said. Those new organizations would just end up relearning lessons the field already knows, according to Talylor. 

“Creating a new charitable institution can take a tremendous amount of time. For donors who want to get to the goal faster, there are existing institutions, like community foundations, that know how to steward this wealth, and nonprofits that need the support now more than ever,” she said. “Let’s use those models instead of waiting and creating.”

Hanna’s says that these potential new funders should back effective existing organizations fully — including staff and overhead, not just programs — so they can grow to match the scale of the problem and build new ones only when nothing else can fill a real gap. She applies the same logic to donor-advised funds, worrying less about whether nonprofits can absorb a grant and more about whether donors actually convert the roughly $250 billion sitting in DAFs into real giving rather than parking it for a tax benefit. 

“How do we make sure people give today, not just someday?” she said, warning that newly wealthy AI employees who wait too long to engage in philanthropy risk having their money funneled toward “the perfect tax deduction” by wealth advisers instead of causes they actually care about.

Goldberg of the Founder’s Pledge agrees that newly wealthy donors will need guidance but not necessarily new institutions from scratch. “We’ve seen again and again that the donors who achieve the most impact are the ones who’ve found the right people to help them live their values: peers and advisers who’ve been in their shoes, got some things wrong, and were willing to learn and grow on their giving journey,” he said.