A new wave of philanthropy is coming. Researchers say to follow the evidence.
Evidence Action
/
AI-related wealth is projected to add $37 billion to $100 billion a year in philanthropic spending over the next few years, a windfall some are now calling the “third wave of American philanthropy”. The Center for Global Development has examined how much of this capital global health and development could absorb effectively, and which uses of it are best supported by evidence.
The result is Putting the AI Windfall to Work, a series that launched in July 2026. The early posts argue for deploying capital at scale on interventions that already work.
In the first post of the series, Rachel Glennerster and Leah R. Rosenzweig contend that $50 billion a year is large but far from unspendable, and point to large-scale implementers, technical intermediaries, and “quick wins” from proven interventions. From there, the series goes deeper on specific problems, with several pieces overlapping with our own work.
Putting the AI Windfall to Work
Selected pieces from the Center for Global Development on where AI-era philanthropy could do the most good.
Kenny ranks vaccines among the world's greatest technologies, noting they have prevented an estimated 154 million deaths since 1974, the great majority among young children, and citing $16 saved in health costs, lost wages and lost productivity for every $1 invested in developing countries. But the poorest people are a weak market, so vaccine research, testing and rollout for their diseases lag. He argues new money could close that gap.
Our work
A vaccine only protects the children it reaches. Nigeria has 2.1 million zero-dose children, 15 percent of the global total. We're preparing a pilot there through our Accelerator for new program development, building on a large randomized trial in India, to test reminders, community ambassadors and incentives that could bring children who have never received a vaccine dose to the clinic.
Bansal opens with the toll: about 2.3 million newborn deaths a year, and, alongside maternal deaths and stillbirths, most of them preventable. Proven tools are underfunded: a PATH report he cites estimates donor aid for maternal, newborn and child health in the highest-burden countries fell 49 percent in 2025. He points to kangaroo mother care, which cuts neonatal mortality by roughly a third, and a $0.84 dual HIV/syphilis test that has reached more than 31 million pregnant women and is estimated to have averted about 90,000 stillbirths.
Our work
Syphilis-Free Start was built to close the gap in antenatal syphilis screening and treatment, which Bansal names as a priority for focused funding.
Kangaroo Mother Care sits within the newborn health work our Accelerator is exploring, with a pilot expected to launch in Liberia in 2027.
Crawfurd argues that service-delivery organizations can only scale so fast, so new money needs more places to go. He makes the case for education on earnings returns to schooling and learning, plus social benefits that are hard to value, and offers a personal, non-exhaustive shortlist of 34 organizations he would bet on. He sees governments as the real route to scale.
“Deworming is still one of the cheapest levers for school attendance that exists.”
A randomized controlled trial in Kenya found that dewormed children had 25 percent less school absenteeism and, two decades later, 13 percent higher incomes.
Researchers and practitioners across the sector do not agree on where new philanthropic capital should go. Some argue for scaling proven interventions, others for paying for breakthroughs that do not yet exist, redesigning how services are delivered, or backing people and movements closer to the problem. The series captures that range of ideas, while our own work belongs to the first of these: interventions with a strong evidence base, a clear cost-effectiveness case, and a realistic path to delivery at government scale.
For donors of any size, we think those are the right questions to ask of any gift. What does the evidence say, what does it cost per outcome, and can it reach the children and families who need it?
Argues that new philanthropists could spend tens of billions a year on proven, cost-effective interventions, pointing to large-scale implementers and technical intermediaries.
Proposes paying for breakthroughs markets will not supply through “pull” funding that pays out only if an innovation works, and building a standing institution to run it.
Weighs funding proven interventions against funding trusted people and movements, and argues for balance with a shift toward people-centered funding, including gender equality and accountable AI.
Argues AI’s biggest gains may come from redesigning how schools and clinics in low- and middle-income countries operate, and that funders should back a portfolio of those bets alongside proven interventions.
