NIH Has Spent 15 Percent of Its Research Budget Halfway Through the Year. NSF Is at 20 Percent of Normal. The Federal Funding Machine Is Broken.

The National Institutes of Health has a $38 billion research budget for fiscal year 2026. As of late March — halfway through the fiscal year — it had obligated $5.8 billion. That is 15 percent.

At the National Science Foundation, the picture is proportionally even worse. The agency has funded 613 grants this fiscal year, roughly one-fifth of the 3,000-plus it would typically fund by this point. The dollar value of awards is running at about one-third of historical averages.

These are not budget cuts. Congress appropriated the money. The agencies have it. They cannot get it out the door.

What is happening across NIH, NSF, and the broader federal research apparatus is not a policy disagreement about how much to spend on science. It is an operational breakdown — a compounding failure of staffing, process, and institutional capacity that has jammed the machinery through which $50 billion in annual research funding is supposed to flow. And the consequences are already cascading through universities, labs, and the careers of thousands of researchers who did everything right and are now waiting for money that may not arrive until it is too late to matter.

Inside NIH's $32 Billion Backlog

The Association of American Medical Colleges published an analysis in late March that quantified the scale of the NIH bottleneck. Through the end of February 2026, NIH had awarded $5.8 billion in external research obligations — compared to nearly $9 billion at the same point in fiscal year 2025. The monthly breakdown reveals the trajectory: $1.2 billion in December, and a combined $2 billion across January and February.

Since October 2025, NIH has awarded only 1,187 new competitive grants — a 63 percent decline from the five-year average for the same period. The agency made 74 percent fewer new competitive awards through early March than the 2021–2024 average. The monetary value of those awards ran 62 percent below historical norms.

Three forces created this bottleneck simultaneously.

The government shutdown. The federal government shut down on October 1, 2025, and did not reopen until November 12 — a historically long 43-day closure at the start of the fiscal year. NIH could not distribute any funds for the first seven weeks. More critically, the shutdown forced the rescheduling of hundreds of grant review panels, the peer-review sessions where scientists evaluate proposals and make funding recommendations. At NSF alone, more than 300 panels had to be rescheduled. At NIH, the disruption was proportionally larger given the agency's scale.

Mass staff losses. NIH has lost thousands of employees through layoffs, early retirements, and voluntary departures over the past year. One NIH institute warned internally that it could leave as much as $500 million in appropriated funding unspent because it simply lacks enough program officers and grants management specialists to process the applications. The grants management workforce — the people who review budgets, negotiate terms, issue awards, and handle compliance — cannot be replaced quickly. Training a grants management specialist takes months. The institutional knowledge lost when experienced staff leave takes years to rebuild.

New review procedures. NIH introduced a "computational text analysis tool" that scans grant proposals for terms the agency deems misaligned with its priorities. While an HHS spokesperson stated that "NIH has never prohibited the use of any particular words in grant applications," researchers report that flagged proposals are delayed while investigators rework language and resubmit. This added review layer, on top of the existing peer-review process, creates friction in a system that was already running far behind schedule.

The result: 14 new Notices of Funding Opportunity in fiscal year 2026, compared to 780 in a typical year. The pipeline of new research calls has essentially stopped, even as billions in appropriated funds sit unspent.

NSF's Parallel Crisis

The National Science Foundation's bottleneck has different origins but similar consequences. NSF entered fiscal year 2026 without a permanent director — Sethuraman Panchanathan resigned in April 2025, and Chief of Staff Brian Stone has been performing director duties since. The agency also absorbed 1,752 grant terminations under DOGE, losing $1.4 billion in committed research funding and the institutional relationships that went with those awards.

Congress did not approve NSF's FY2026 budget until late January 2026 — nearly four months into the fiscal year. Combined with the fall shutdown, NSF spent nearly half the fiscal year operating under continuing resolution authority, unable to issue new multi-year awards or launch new programs.

Then came the merit review overhaul. In December 2025, NSF quietly reduced the minimum number of external peer reviews per proposal from three to two, allowed one review to be conducted by NSF staff, made panel discussions optional, and compressed panel summaries to three to five sentences. The agency cited "proposal backlog and workforce reductions" — an acknowledgment that it lacked the capacity to maintain its own quality standards.

The 613 grants funded through mid-April represent just the proposals that cleared every hurdle: peer review, program officer recommendation, budget negotiation, and grants management processing. Thousands more sit in various stages of the pipeline, waiting for staff who are not there, reviews that have not been scheduled, and decisions that have not been made.

The Year-End Funding Dump Problem

Federal agencies operate on an October-to-September fiscal year. Funds appropriated for FY2026 that are not obligated by September 30 do not simply carry over — for many account types, they expire. This creates enormous pressure to push money out the door in the final quarter.

It happened in FY2025. NIH obligated more than 50 percent of its necessary research funds in the final three months of the fiscal year, relying heavily on controversial multi-year "lump sum" grants that changed the fundamental structure of how awards are distributed. Fewer researchers received awards, but each award was larger and covered multiple years — a structure that reduced the total number of funded investigators while technically spending down the appropriation.

Fiscal year 2026 is tracking toward the same pattern, but with a larger backlog and fewer staff to process it. If NIH needs to obligate roughly $32 billion in the remaining five months of the fiscal year, it will face an impossible throughput problem. The likely outcome is another round of lump-sum multi-year awards — concentrating funding in fewer hands, reducing the number of new investigators who receive any support, and creating the appearance of spending without the reality of broadly distributed research investment.

For researchers waiting on pending applications, this creates a paradox. The money exists. The proposals have been reviewed. But the award may not come until August or September — and when it does, it may look different from what was proposed. Multi-year lump-sum structures change lab planning, hiring timelines, and research trajectories in ways that quarterly-funded awards do not.

Universities Are Already Responding

The downstream effects of the funding bottleneck are not theoretical. The Association of American Universities reported that member institutions have implemented hiring freezes and reduced PhD admissions in the life sciences in direct response to NIH funding uncertainty.

Early-career researchers are bearing the sharpest impact. NIH R01 grants — the gold standard for independent investigators — went to 1,114 applicants in fiscal year 2025, down from 1,423 in fiscal year 2024. The applicant pool actually grew, from 5,446 to 6,065, meaning more researchers competed for fewer awards. Success rates are plummeting at exactly the career stage when scientists are most vulnerable to funding disruptions.

University research offices report that faculty are submitting proposals into what feels like a void. Applications go in, acknowledgments come back, and then silence — for months. The lack of communication compounds the uncertainty. Researchers cannot tell whether their proposal is in a review queue, stuck in the text-analysis screening step, waiting for a panel that has not been rescheduled, or simply lost in an understaffed grants management office.

What Grant Seekers Should Do in the Bottleneck

Assume delays and plan accordingly. If you have a pending NIH or NSF application, build your lab budget and personnel plans around a scenario where the award arrives in the last quarter of the fiscal year — or not at all in FY2026. Do not make hiring commitments based on expected award dates. Bridge funding from your institution, if available, buys time that assumptions about federal timelines do not.

Contact program officers directly. Pre-submission communication with program officers has always been good practice. In the current environment, it is essential. Program officers can tell you whether a solicitation is actively reviewing, whether panels have been scheduled, and whether your research area is likely to receive new awards this fiscal year. At NSF, this is now critical given the reduced review infrastructure.

Diversify across agencies and mechanisms. The Department of Defense, the Department of Energy, ARPA-H, and USDA all fund research that overlaps with NIH and NSF portfolios. SBIR and STTR programs have just reopened after a six-month lapse, with agencies releasing new topics across DoD, NASA, NIH, and NSF simultaneously. Private foundations are increasing payouts to compensate for federal disruptions.

Watch for lump-sum structural changes. If your award arrives as a multi-year lump sum rather than annual increments, the budget management requirements are fundamentally different. You receive years of funding at once, with institutional overhead taken upfront and spending flexibility that comes with different reporting obligations. Plan for this possibility now rather than being surprised by it in September.

Document everything for resubmission. If your application is delayed or declined, preserve all correspondence, reviewer feedback, and program officer communications. The current review environment — with fewer reviewers, compressed summaries, and computational screening — produces less feedback than historical norms. What you receive is more valuable and harder to replace.

The Structural Problem Behind the Bottleneck

Budget debates will continue. Administrations will propose cuts, and Congress will negotiate. But the bottleneck problem is distinct from the budget problem. Even if Congress appropriates full funding for FY2027, the agencies lack the staffing and operational capacity to distribute it efficiently. The grants management workforce has been depleted. The review infrastructure has been degraded. The institutional knowledge that kept the system running has walked out the door.

Rebuilding that capacity will take years and sustained investment in the people and processes that convert appropriations into research awards. Until then, the federal research funding system will operate well below its theoretical capacity — a machine with fuel in the tank and a broken transmission.

For researchers navigating this broken pipeline, tools like Granted can help identify which funding streams are actually flowing, match your work to programs across multiple agencies, and build proposals that survive in a review environment where fewer eyes see each application and every word matters more than it used to.