The 205 Phantom Forecasts: How NIH Is Withholding Research Funding Through Process, Not Budget

The most consequential federal grant story of 2026 is not a budget cut. It is a queue.

In early May, NPR published an analysis by Elizabeth Ginexi, a former 22-year program officer at the National Institutes of Health, that quantified something the biomedical research community had been describing anecdotally for months. Of 336 NIH funding forecasts still publicly listed as open opportunities, 205 had blown past their promised posting dates with no full announcement ever published. Sixty-one percent of the agency's advertised research priorities, by Ginexi's count, exist only on paper. They appear in NIH's forecasting system, they show up in institutional grants dashboards, they tell scientists what to write toward — and they never convert into solicitations that researchers can actually submit applications against.

The phantom forecast finding is one data point. It connects to several others that, taken together, describe a research funding system that is being slowed and constrained without a single appropriations vote. NIH had awarded 1,385 competitive grants in fiscal 2026 as of March 31 — down 54 percent from the 3,024 awarded by the same date in fiscal 2025, and 74 percent below the 2021–2024 average for the same period. The agency cannot freely spend the budget Congress appropriated in February. It must submit detailed "spend plans" to the White House Office of Management and Budget for approval before obligating research funds. A new policy mandates that multi-year grants be funded with a single upfront lump sum rather than year-by-year, an accounting change that has crashed funding rates at the National Cancer Institute from one in ten applicants to one in twenty-five. In August 2025, the Government Accountability Office formally determined that the Department of Health and Human Services had violated the Impoundment Control Act by withholding NIH funds from obligation and expenditure.

This is the architecture of a slowdown that did not require a single dollar to be rescinded.

What a "forecast" is supposed to do

To understand why 205 phantom listings matter, it helps to understand what NIH forecasts are for. The forecasting system is the agency's signal mechanism to the research community. Months before a Notice of Funding Opportunity goes live, NIH posts a forecast describing the topic, expected scope, anticipated dollar range, and target publication date. A principal investigator at a research university reads those forecasts and uses them to decide what direction to take her next pilot study, which postdoc to hire, whether to apply for institutional bridge funding to cover a gap, and how to scope a sabbatical year. University grants offices read them to staff up internal review committees. Department chairs use them to project budget for the next academic year.

When a forecast posts, the lab work behind it begins immediately. A researcher does not wait until the formal announcement to start drafting Specific Aims and recruiting collaborators. The forecast is the actionable signal. The announcement, when it comes, is the deadline.

The Ginexi analysis means that 205 of those signals fired without the deadline ever arriving. Labs reoriented. Postdocs were recruited against specific anticipated programs. Pilot data was collected. And then the announcement that would have allowed an actual application simply did not appear. The labs did not get told the program was cancelled. They got told, by silence, that the program had been put on indefinite hold.

This is qualitatively different from a budget cut. A cut is a decision. A phantom forecast is a non-decision indefinitely sustained. There is nothing for an applicant to appeal, no advocacy group to lobby, no political cycle to wait out. The funding direction was announced; the funding never arrived; the institutional damage compounded across hundreds of labs simultaneously.

How the slowdown is engineered

The phantom forecasts sit downstream of three operational mechanisms that together explain the 54 percent year-over-year drop in NIH competitive awards.

The first is the spend-plan requirement. NIH's fiscal 2026 budget was signed into law on February 3, 2026. Under normal practice, the agency would obligate against that budget on a rolling basis through the fiscal year, with institute directors and program officers making award decisions on the timetable that peer review and council cycles dictate. Under the current regime, NIH must instead submit and receive OMB approval for detailed spend plans before obligating research funds. Those plans cover not just headline allocations but program-by-program funding intentions. Each round-trip with OMB adds weeks. Each program with disputed scope can stall indefinitely while the agency rewrites narratives to fit administration priorities. The result is a system in which appropriations have been approved but expenditures have not been authorized.

The second is the lump-sum multi-year policy. Historically, NIH awarded multi-year R01 grants on a year-by-year obligation basis: the institute committed to five years of funding in principle but obligated only the current year's tranche, freeing the remaining commitment to come out of future-year budgets. The new policy requires that the full multi-year commitment be obligated upfront against the current fiscal year's appropriation. The arithmetic is brutal. An institute that previously could fund 100 R01s with $40 million in annual obligations now needs $200 million up front to fund the same number of five-year grants. Holding budget constant, success rates collapse. The National Cancer Institute's drop from one-in-ten to one-in-twenty-five is the most cited example, but the mechanism applies across institutes.

The third is the additional review layer. The OMB director publicly acknowledged that his office is reassessing already peer-reviewed grants — a step that historically would have been considered scientific interference and that adds weeks or months of delay to grants that have already cleared study section and council. Each layer of review is a queue, and queues compound. By the time a grant has cleared peer review, council, institute prioritization, spend-plan approval, and OMB reassessment, the calendar has advanced past the point where the originally-targeted budget cycle is still available.

The phantom forecasts are what this system produces at the front end of the pipeline. Programs that were on schedule in the forecast layer cannot make it through to the announcement layer because the institutes do not have approval to commit the funds those announcements would require.

The downstream collapse

Several large patterns flow from the slowdown that are not visible if you only look at the headline appropriations number.

Federal grant opportunities listed across all of grants.gov fell from roughly 2,400 in early 2025 to about 1,600 by February 2026 — a one-third decline that reflects withholding patterns far broader than NIH alone. Posted grants currently accepting applications dropped from 1,850 to roughly 900, a 50 percent contraction. The pipeline of work that the federal grant ecosystem ordinarily generates for principal investigators, contract research organizations, technology transfer offices, and university research administration has been cut nearly in half over fifteen months without any single legislative action visible to the public.

Compounding this, application windows that remain open have compressed. Grants that historically would be open for six months are now closing in weeks. For institutions accustomed to building consortia, securing letters of support, and running internal limited-submission competitions on a six-month timeline, the new compressed cycle simply rules out competing for awards that require collaborative architecture.

The cumulative effect on early-career scientists is the most under-discussed component. A first-year assistant professor with a four-year tenure clock cannot wait two years for the NIH pipeline to clear before securing R01 funding. The lab is hiring postdocs, ordering equipment, and committing to recruitment cycles against the assumption that the federal research economy is operating on something close to its normal cadence. A 54 percent drop in award volume in a single fiscal year is not a slowdown that early-career scientists' career timelines can absorb. The exit decisions being made in spring and summer 2026 — to industry, to other countries, to careers outside research entirely — will not show up in federal statistics for years.

What this means strategically

For PIs, university research offices, and the broader research community, the operational implication is that NIH forecasts can no longer be treated as actionable signals. A forecast posting in 2026 is not a commitment that an announcement will follow. Institutions that staff up against forecasts — bringing on postdocs, signing equipment leases, committing institutional bridge funding — are accumulating financial exposure against a pipeline that may never materialize. The new prudent default is to treat forecasts as aspirational until an announcement actually appears in the NIH Guide, and to scope lab planning to current open NOFOs rather than the forecast inventory.

For grant strategists advising biomedical research applicants, the priority targets in the 2026–2027 window are programs that have already converted from forecast to announcement and have funding committed against current-year spend plans. The signal of a healthy program is not the existence of a forecast — it is the existence of an open announcement with applications currently being reviewed. Cross-reference NIH RePORTER for evidence that the institute is actively making awards against the program. Programs where the most recent award predates the new fiscal year are likely paused, regardless of what the forecast page says.

For the broader nonprofit and research advocacy community, the GAO Impoundment Control Act ruling from August 2025 is the most important legal predicate available. The administration has been formally found to be withholding funds Congress appropriated. That finding does not automatically release the funds, but it provides litigation footing and congressional oversight leverage that has not been fully deployed. The 64 lawsuits already filed challenging federal research funding actions are running on parallel tracks; the Impoundment Control Act violation is the procedural anchor underneath several of them.

For applicants outside biomedical research, the lesson generalizes. The pattern Ginexi documented at NIH — forecasts that exist on paper, announcements that never publish — is the same pattern researchers are reporting at NSF, the Department of Energy Office of Science, and across the federal grant opportunities that dropped 33 percent in early 2026. Phantom listings are now a feature of the federal grant landscape, not an anomaly.

The honest read on the 2026 federal funding environment is that the budget has not been cut in the dramatic way the early-year political coverage suggested. Congress appropriated. The president signed. The money was not, however, allowed to move. The 205 phantom forecasts are the visible surface of a much deeper hold on obligation, and they will continue to be the leading indicator for whether the science funding pipeline reopens in the second half of the fiscal year — or whether the slowdown becomes the new operating equilibrium.

For grant applicants planning the next twelve months: assume the slowdown holds, scope to current open announcements rather than forecasts, and build state, foundation, and industry pipelines as primary rather than supplementary revenue. The federal research economy is not closed, but it is operating at half its historical throughput, and there is no signal yet that the pace will recover before the FY2027 appropriations cycle begins.