Federal Grants to Everyday Charities Fell 38% in Trump’s First Months

The analysis was conducted by Granted AI, a company that uses AI to help nonprofits search for grants. It strips out hospitals and universities as well as spending sprees on rural electric cooperative and clean-energy projects, to try to get a handle on how much funding was cut to the everyday charities that many view as the core of the nonprofit sector.

The grants to these charities fell from $35.4 billion in 2024 to $21.4 billion in 2025 during the eight months from the beginning of February, the first full month Trump was in office, through the end of the government’s fiscal year in September. The decline puts a hard number behind the turmoil caused by repeated attacks on nonprofit programs that made headlines throughout Trump’s first year in office. The administration dismantled the U.S. Agency for International Development, then abruptly canceled grants to criminal-justice groups, cut funding to arts and humanities organizations, and canceled funding for NPR and PBS, among other major shifts in policy affecting grants to charities.

Nonprofits have long relied on federal funding to help them provide vital services to their communities. Those funds were often viewed as more reliable than funding from grant makers, which can shift priorities, and donors whose interests can be fickle. The sudden loss of federal dollars has upended finances for charities across the country in ways that are still ricocheting through the nonprofit world, Bergdoll said. When faced with funding cuts and spiking needs, charities are likely to prioritize their programmatic work over their operations. That may prompt cuts in administrative roles — including fundraising — and could lead to even more severe long-term financial challenges, he said.

“We certainly are seeing some impacts immediately,” Bergdoll said. “I think other impacts will take longer to be felt.”

Asked via email to comment on the findings, the White House referred the Chronicle to the State Department, which did not respond to multiple emails.

Using AI to Parse Data

The Chronicle was unable to find any similar analysis of federal grant making data over this period of profound change. Granted AI, which used AI to analyze the data, is a grants database primarily used by nonprofit fundraisers, some of whom pay fees for advanced services. It was started by Nathan Eagle in 2021 after his sister-in-law, who runs a youth science program in Cottonwood, Ariz., reached out to him for help finding new grants.

Eagle has an impressive academic and business background. He earned a Ph.D. from MIT Media Lab, started a company to bring mobile internet to developing countries, and was a finalist for a MacArthur Fellowship, widely known as a “genius” grant.

For his research on federal grants, Eagle relies on a resource that’s long been available to the public — USAspending.gov, which has comprehensive databases of federal spending, including grants. Eagle worked alone but employed six AI agents to collect 7 million award transactions, crunch the data, and then check each other’s work.

In stripping out universities, hospitals, and others, Eagle is narrowing this analysis to everyday charities and measuring only a small slice of the government money that flows to nonprofits. In February 2025, Candid conducted an analysis of federal Form 990 informational tax filings by more than 100,000 charities and found that they received a total of $303 billion per year from federal, state, and local governments — many times greater than the spending that Granted AI captures in its narrower evaluation. The Candid total includes federal spending that flows through states before being disbursed to nonprofits. That spending isn’t included in Granted AI’s analysis.

Eagle said he has “very high confidence” in his research. He openly shares the findings and invites feedback from others.

“As you probably can tell, I drink the AI Kool-Aid,” Eagle said. “When you start putting agents on this, they are just really good at being able to fix data corruptions, get it cleaned in the right way, and make it actionable in ways that I think historically have been hard.”

Federal Funding Can Be Hard to Track

In compiling the statistics on direct federal grants to charities, Eagle is wading into waters that other charity researchers have generally stayed away from. The last report on direct federal grants to charities came from the U.S. Government Accountability Office in 2009 and was based on 2006 data, Eagle said.

Researchers from the Urban Institute, National Council of Nonprofits, and Partnership for Public Service all agreed that tracking direct federal funding to charities is extremely challenging. Not all nonprofits listed as charities in the USAspending.gov data are actual 501(c)(3)s, experts say. Many federal grants are missed because, as noted earlier, they flow to states first before ending up at charities — and states don’t have consistent reporting mechanisms. And some grants are hard to track because they run through nonprofit intermediaries — including universities — before ending up at smaller nonprofits.

While those challenges remain for Granted AI, just like any other research organization, some experts say AI can still streamline the data and evaluation process — and make federal data more useful.

“I suspect where AI has been super helpful for Nathan is in building a data workflow to analyze things,” something that has taken an army of assistants in the past, said Jesse Lecy, director of the National Center for Education Statistics. “The fact that he’s doing it all on his own without prior expertise in this area says a lot about some of the potentials of AI.”