Grant Witness

Grant Witness[^1] is a project to track the termination of grants of scientific research agencies under the Trump administration in 2025. We currently are tracking terminations of grants from the National Institutes of Health (NIH) and the National Science Foundation (NSF).

Our data on terminated NIH and NSF grants are collected from submissions from affected researchers as well as government websites and databases. We encourage researchers, program officers, and grant administrators to submit information via the forms here (NIH) and here (NSF) to help us keep our data up to date.

Grant Witness is a project of Noam Ross, Scott Delaney, Anthony Barente, Emma Mairson, Eric Scott, and Mally Shan with input and support from additional volunteers.

For tips, questions, corrections, or press inquiries, contact us at info@grant-witness.us. You may also reach us on Signal at sdelaney.84.

Resources

Visualizations

For more information, use your cursor to click on or hover over a state. The below button allows you to download a static .png version of the map.

For more information, use your cursor to click on or hover over a state. The below button allows you to download a static .png version of the map.

Legend

The map shades states based on the number of grants that were ever impacted, which includes ‘terminated’ and/or ‘frozen’ grants. Additionally, even if a grant is listed as ‘possibly reinstated’, it is included in the total grants that were ever impacted, as it has at one point been ‘terminated’ and/or ‘frozen’. A percentile ranking method was employed to determine the impact of ‘terminated’ and/or ‘frozen’ grants in a particular state relative to the grants impacted in other states. This method was chosen due to the large range in the number of impacted grants across the country and to ensure consistency in shading for the map.

Number of Grants Ever Cut (or Frozen)

When hovering/clicking on a state, you can see the current number of grants that are currently ‘terminated’ (and/or ‘frozen’ for the NIH data) for that particular state. This number does include any grants that are ‘possibly reinstated’, since it has at one point been ‘terminated’ (and/or ‘frozen’ for the NIH data).

Number of Grants Currently Cut (or Frozen)

When hovering/clicking on a state, you can see the current number of grants that are currently ‘terminated’ (and/or ‘frozen’ for the NIH data) for that particular state. This number does not include any grants that are ‘possibly reinstated’.

Number of Grants Reinstated

When hovering/clicking on a state, you can see the current number of grants that are ‘possibly reinstated’ for that particular state, as identified through the ‘status’ variable in the respective data sets.

Estimated Current Loss

When hovering/clicking on a state, you can see the estimated current loss in grant funding for that particular state. This value represents the total amount of funds remaining in grants that are currently ‘terminated’ (and/or ‘frozen’ for the NIH data); ‘possibly reinstated’ grants are not included in this total. - For the NIH data, the total amount of remaining funding is calculated as the sum of “total estimated remaining” (calculated as the ‘total award’ minus the ‘total estimated outlays’).

  • For the NSF data, this value is calculated as the sum of “estimated remaining”, which is the unspent funds within an already obligated amount for a grant as tracked on USAspending.gov.

Methods

NIH Methods

Data in our database of terminated NIH grants come from several sources. Among them are affected Principal Investigators (i.e., scientists), who we encourage to submit information on their terminated grants using a Google Form, which we have designed and published specifically for this purpose. We also aggregate data from news reports, social media, Doge.gov, USASpending.gov, NIH’s X feed, NIH RePORTER, and the HHS TAGGS system. The HHS TAGGS system provides data in two ways. First, it details grant-based financial transactions with information that is updated at least daily. Second, TAGGS administrators periodically (~1/week) release a PDF list of terminated grants. The TAGGS PDF list is one source of information in our database among several. Because we aggregate from multiple sources, our database includes grants not yet listed on federal government websites, including the HHS TAGGS PDF of terminated grants, and we believe it is the most comprehensive, up-to-date resource currently available that quantifies and characterizes terminated NIH grants.

Estimating NIH grant budgets

Determining the overall grant size, total funds outlaid, and total unspent funds for each terminated NIH grant is a complex process due to NIH’s system of separate tracking of grant years, supplements, and subawards, all under a single Federal Award Identification Number (FAIN). For top-level or “parent” grants, we use data from USAspending.gov, which reports to total obligated funds for the full multi-year grant, and the total outlays. USAspending.gov data is typically 1-2 months behind current data, and at the time of termination, will not include funds spent for close-out. For supplements and subawards, USAspending.gov does not break out budgets. For a subset of supplements and subawards, we use totals and outlays as reported by HHS in the PDF of terminated grants posted to their website. As only a subset of terminated grants are included in that file, for remaining supplements and subawards we use the total supplement/subaward budget for the current grant year as listed in NIH RePORTER, and estimate the total outlays based on the percentage of the grant year that has elapsed. This assumes a constant rate of spending, and under-estimates the total budget and outlays for multi-year supplements and subawards. For all awards and data sources, we calculate remaining unspent funds as the difference between the total budget and total outlays.

Flagged words

NIH grant records contain a flagged_words field. Terms in this field are those that appear in the the grant’s title, abstract, or public health relevance statement and match a list of words the the administration has reportedly used to identify grants for termination. The full list of words is drawn from this New York Times article and supplemented with variants of the words “vaccine”, “hesitancy”, and “mRNA”. We do not know specifically whether these words are used to classify grants for termination, but include them for further analysis. Some flagged words (e.g., “bias” or “trans”) have multiple meanings in scientific literature, and may be matched to parts of words (e.g., “transcription”).

Grant Topics

We highlight groups of grants under specific topics of interest as tabs in the NIH table. These topics range from research subject areas targeted by the administration (e.g., Transgender and LGBTQ health, Vaccine Hesitancy), diseases of interest (e.g., HIV, Cancer, Alzheimer’s), research activities (e.g., Training), or institutions (targeted Universities such as Columbia and Harvard, Minority-serving institutions). Each of these topics is implemented as a filter making use of fields in the data such as flagged words, NIH-assigned terms, or institutional traits. After clicking on a tab, clicking the “Filter” field will show the criteria for the topic.

Grants are classified on different tabs using multiple criteria. We identify grants investigating specific topics (e.g., LGBTQ health or “DEI” topics) based on the presence of “flagged words” that may appear in each grant’s title, abstract text, public health relevance statement, or project terms as they appear on NIH RePORTER. “Flagged words” are those previously identified by the New York Times (link below) as words the current administration seeks to purge from Federal programs. Some flagged words (e.g., “bias” or “trans”) have multiple meanings in scientific literature. To reduce misclassification, we use a subset of the NYT flagged words list to classify grants based by subject matter. The complete list of flagged words is here: https://www.nytimes.com/interactive/2025/03/07/us/trump-federal-agencies-websites-words-dei.html.

Frozen grant funding

For more information about how we track frozen grants, see here.

Reinstated grants

For more information about how we track court-ordered grant reinstatements, see here.

NSF Methods

We currently aggregate terminated NSF grant information from Principal Investigators and other affected scientists, as well as from other user-submitted lists of terminated grants. We augment this data with information from NSF’s Award Search and USAspending.gov. Beginning 2025-05-22, NSF released a list of terminated grants, which we incorporate even as we monitor additional sources. Previous experience suggests official lists may be inaccurate and/or out-of-date (The 2025-05-22 official NSF list included 90+ grants terminated after their end date).

Estimating NSF grant budgets

For NSF awards, a single grant typically matches 1-to-1 with a a single Federal Award Identification Number (FAIN). We report both NSF’s total expected budget for the grant, and funds obligated and outlaid as reported by USAspending.gov. For “continuing” grants, funds are obligated incrementally over multiple years, so the total expected budget is the better representation of the full value of the grant. We also include the total funds outlaid as reported by USAspending.gov, and calculate the remaining unspent funds as the difference between the total expected budget and total outlays. However, USAspending.gov data is typically 1-2 months behind current data, and we find it is more likely to be under-reported for NSF than NIH.

EPA Methods (Preliminary)

EPA grant terminations are collected from EPA project managers, as well as search of the EPA Next Generation Grant System (NGGS) and to identify terminated grants via changes in grant dates and deobligations. NGGS does not include all EPA grants. We include total award value (sum of pre-termination obligations), outlaid funds, and reamining unspent frunds from USAspending.gov. As above, these numbers may be 1-2 months behind current data. For EPA grants, we also include a timeline of grant amendments and outlays in the current year. These provide more tranparency on events determining grant termination and reinstatement status.

Transparency

To ensure each reader can objectively verify all details of our database, we list NIH- and NSF-issued grant serial numbers, which are unique to each grant, website hyperlinks to federal databases describing each grant, and many other details for each terminated grant. These efforts to ensure transparency are critical for multiple reasons. Foremost is that the NIH and NSF grant termination processes have been extraordinarily chaotic—far more so than is often reported in news articles. In many cases, the government has often provided incorrect and shifting information (knowingly or not) about exactly which grants were terminated, when, and why. Their estimates of the dollar value of grants terminated are even more unreliable. We also believe the government has restored a limited number of NIH grants that it previously terminated, often without providing public notice. As a result, our database may contain a small number of minor inaccuracies. To report corrections or ask questions, please contact Scott Delaney on Signal at sdelaney.84. [^1]: Grant Witness was previously called ‘Grant Watch’ but is not, and never has been, affiliated with grantwatch.com