In early 2025, as reports of mass grant terminations at NIH and NSF rippled across the scientific community, two researchers stepped in to provide what was missing: a comprehensive, verifiable record of what was being lost. Noam Ross, an ecologist and data scientist, began digging through government databases to see whether the terminations could be tracked. Scott Delaney, an epidemiologist and lawyer, opened a public spreadsheet to crowdsource reports from affected scientists. Their efforts soon merged into what is now the most comprehensive public record of science funding terminations in the United States: Grant Watch.
Grant Watch is a volunteer-run effort to track the termination of federal research grants, specifically those issued by the National Institutes of Health (NIH) and the National Science Foundation (NSF). Since its inception, the project has compiled data on nearly 4,000 grants. According to Grant Watch’s most recent weekly report, NIH alone accounts for approximately $3.2 billion in ‘lost funding’—defined as funding awarded but not paid out due to terminations. However, Ross estimates that across both NIH and NSF, the total value of terminated grants exceeds $10 billion, including both already-spent funds and future support that will no longer be realized. Roughly half of that amount represents sunk costs for work already begun; the other half consists of future funding that will now never materialize.
The economic value of basic research is substantial: federal science funding consistently yields returns that far exceed its costs. In that sense, it is one of the rare public investments that effectively pays for itself. If the current budget proposal is enacted, the estimated $10 billion in terminated grants would translate into significantly greater losses in future productivity, innovation, and economic growth.
From Spreadsheets to Federal Court
The scope of the terminations is unprecedented, and that is what prompted Ross, Delaney and their colleagues to start collecting data. “We sort of stumbled into doing this. Scott and I, and not just us, each started something independently and somewhat joined forces. Suddenly we have a code base and a website, a growing team and a Signal hotline, and now it was cited by a federal judge from the bench.”
What began as spreadsheets and amateur sleuthing now underpins lawsuits from organizations like the ACLU, feeds investigative reporting by national news outlets like the Atlantic and the New York Times, and has been cited in federal court filings. The team—which also includes data scientists Anthony Barente and Emma Mairson and a growing team of volunteers—gathers information from public databases, whistleblowers, and grant recipients themselves, who can submit documentation through web forms. They vet each case, verify grant IDs and award data, and attempt to match submissions with official government records.
The group operates without institutional funding. “We are in some conversations with funders to just get enough to bring in someone for a day and a half a week to help with back-end database and website maintenance stuff, so we can focus on more investigations and outreach. But so far no one has paid for anything,” Ross said.
Ross, who also works on infectious disease forecasting and serves as executive director of rOpenSci, a project that focuses on improving the R programming language, sees Grant Watch as a kind of data journalism project—one that enables rather than replaces traditional reporting. “We can’t tell all the stories.We try to present the information as richly and with as much context as possible so that others can.”
That model is working. In just a few months, Grant Watch has become an essential resource for understanding the rapid dismantling of American scientific infrastructure. While official agency reports are often incomplete or internally contradictory, Grant Watch cross-references its data, flags discrepancies, and updates entries as new information becomes available.
When asked why anyone should trust the data that Grant Watch is producing, Ross replied: “My first and favorite reason is that scientists are very quick to tell us if we are wrong about their stuff. We have made mistakes, but we’ve been told immediately about them every time.”
Why Transparency Matters
Despite its technical complexity, the project is ultimately motivated by a basic civic concern: transparency. The federal government is not just reducing science budgets; it is actively terminating grants, in some cases midstream, without clear justification or process. The result is not only the loss of valuable research but also a profound disruption to the scientific enterprise itself—students without funding, labs disbanded, data lost.
And while the political motivations remain murky, the operational reality is plain to see. As Ross put it: “We can’t tell all the stories. But we can make sure no one can say they didn’t know.”
