One component carried almost the entire dollar effect
GAO found that four Department of Homeland Security components terminated 362 grants between January 20 and September 30, 2025 and deobligated $1.001777 billion. FEMA accounted for $999.433 million—99.77% of that total.
FEMA’s 215 terminated grants represented 59.4% of the count. U.S. Citizenship and Immigration Services terminated 111 grants but accounted for $2.342 million; the Science and Technology Directorate’s 35 grants accounted for $2,000, and CISA’s single termination recorded no deobligation.
| Component | Terminated grants | Deobligated |
|---|---|---|
| FEMA | 215 | $999.433M |
| USCIS | 111 | $2.342M |
| Science and Technology | 35 | $2,000 |
| CISA | 1 | $0 |
| DHS total | 362 | $1.001777B |
The February pause disrupted projects before some actions were reversed
DHS paused disbursement of all obligated grant funding in February 2025 while it carried out a 30-day review. GAO found that some disaster-mitigation projects were delayed nearly a year and that DHS did not fully achieve planned cost-reduction actions.
FEMA reallocated certain terrorism-prevention funds away from some states, then reinstated them after a court order. That does not mean every termination was unlawful or every dollar was ultimately restored. It shows the first process produced reversals and program delays that DHS had not yet systematically incorporated into later decisions.
The original finding is concentration, not a $1 billion fraud claim
The public table can look like a department-wide savings result. Recalculation shows the financial effect was overwhelmingly a FEMA event. That matters because disaster preparedness and mitigation have statutory purposes that differ from immigration or research grants.
Deobligation removes spending authority from an award; it is not evidence the grantee committed fraud, and it is not automatically a permanent Treasury saving if funds are later restored, reallocated or spent through another route.
Method, response and limitations
The newsroom divided each component’s published deobligation by the DHS total and calculated grant-count shares from GAO’s table. Dollar shares are rounded after calculation, and GAO’s source combined DHS, USAspending and agency information.
GAO recommended a process requiring grant-making components to apply lessons from earlier pauses and terminations before future actions. DHS agreed; the recommendation was open. The summary does not provide a final project-by-project ledger, so this report does not total later restorations.