Research Funding in 2026: How the Money Moves, Who Decides, and What Professionals Should Do Next

Research funding is usually treated as plumbing — assumed to work, noticed only when it fails. Between 2025 and 2026 it failed loudly enough to become a strategic problem for deans, finance officers, R&D directors, trial coordinators and early-career scientists.

Three things shifted at once: the gap between budget requests and money actually reaching laboratories widened; the mechanics of funding (overhead reimbursement, obligation timing, award terms) moved from back office to front page; and the global distribution of spending changed, with global R&D estimated at around USD 3.8 trillion in 2024, of which OECD countries accounted for almost two-thirds. OECD

This article explains how research funding is structured, what has actually changed as distinct from what has been proposed, and what professionals can reasonably do about it.

---

Executive summary

• The US FY2026 request proposed roughly $27.5 billion for NIH, about a 40% cut below 2025, with 19 institutes consolidated into eight. Congress rejected it: the enacted bill provided $47.2 billion, about 0.9% above FY2025. BrookingsAmerican College of Radiology
• Appropriation ≠ disbursement. Halfway through FY2026, NIH had obligated $5.8 billion of extramural funding — roughly 15% of the estimated $38 billion due by 30 September. AAMC
• The First Circuit unanimously held NIH could not impose an across-the-board 15% overhead cap, and no Supreme Court petition was filed within the 90-day window. Higher Ed DiveACS C&EN
• OMB's proposed rule of 29 May 2026 would expand political appointees' role in grant approval and ease post-award termination across more than $1 trillion in annual federal funding. American Association for Cancer Research
• The EU proposed nearly doubling its framework programme to €175 billion for 2028–2034 within a €409 billion Competitiveness Fund. europa
• China's R&D reached about $549.7 billion in 2025, with intensity at 2.80% of GDP — above the OECD member average for the first time. China Data Portal
• Philanthropy is catalytic, not substitutive: foundations and non-profits supplied about 15% of US basic and applied research funding at universities and non-profits in 2023, or roughly 21% including endowment-derived support. PHYSICS TODAY

---

The architecture: where research money comes from

Table 1. Funding sources compared. Original OneWise analysis.

Institutional self-funding is larger than most assume: universities self-funded $19.2 billion of research expenditure in 2023, or 17% of total support. Philanthropy is structurally small — foundations, associations and societies provide roughly 1.2% of medical and health research funding against about 25% from the federal government. SciphilMedscape

Direct vs indirect costs. Direct costs attach to a project (salaries, reagents). Indirect — facilities and administrative (F&A) — costs cover buildings, utilities, biosafety, compliance and data security, historically reimbursed at institution-specific negotiated rates.

Callout: Overheads are neither "bloat" nor purely infrastructural. Rates reflect audited cost structures, so a hospital-based cancer centre legitimately differs from a small college — but they also embed cross-subsidy decisions rarely disclosed. The reform debate is about transparency as much as the number.

---

What actually changed in 2025–2026

FIGURE 1 — "The Three-Gate Funding Pipeline." Purpose: show why a healthy appropriation doesn't guarantee a funded lab. Components: horizontal flow, Request → Appropriation → Obligation → Laboratory, with three gates. Labels beneath each gate: "Signals priorities; not binding" / "Legally binding topline" / "Timing, terms, terminations." Arrows: solid left-to-right; dashed drops to a risk icon at each gate. Hierarchy: Gate 3 largest. Caption: "Most 2026 funding risk sat at the obligation stage, not the appropriation stage."

Table 2. Key events. Compiled by OneWise.

P.L. 119-75 was signed on 3 February 2026. For NSF, reported FY2026 figures vary by account basis — one summary records $9 billion against $9.06 billion in FY2025, while trade reporting cites $8.8 billion — but all sit far above the $3.9 billion request. Cite the basis when quoting these numbers. Congress.gov + 3

The overhead ruling. NIH announced the 15% cap in a notice of fewer than three pages, effective the next business day; it expected $4 billion in savings. The First Circuit affirmed jurisdiction and found the action violated statute and regulation. Two caveats: the ruling turned substantially on an appropriations rider, renewed annually; and reform continues — the Joint Associations Group's FAIR model would differentiate by institution and research type rather than negotiating university-wide rates, with FY2026 bills including language to advance it. First Circuit + 4

Rewriting the rules. Despite a 45-day window, nearly 500,000 comments were filed on the OMB proposal, and an analysis of more than 52,000 public comments found approximately 95% opposed; OMB intends to finalise by 1 October 2026, while the August continuing resolution would push it to 11 December. Separately, the Department of Energy acknowledged in federal court that $7.6 billion in clean-energy grants were selected for cancellation based solely on the political identity of the recipient's state. "Award received" can no longer be modelled as terminal. American Association for Cancer ResearchAmerican Association for Cancer Research

---

Allocation: how funders choose

Table 3. Allocation mechanisms. Original OneWise synthesis.

Researchers describe conditions where allocation can consume more resources than it distributes — the Szilard point — and note peer review's documented bias against unconventional or risky proposals. Reported strain is significant: one analysis reports UKRI competitive applications nearly doubling since 2017 while award rates fell from 36% to 19% — indicative only, as it derives from trade reporting cited in a preprint. niharxiv

New Zealand's Health Research Council became the first major government funder to allocate by lottery, for Explorer Grants. In the applicant survey, most said randomisation changed neither their approach (69%) nor their preparation time (75%), with support notably higher when limited to applications of comparable quality. More recently, a German funding line found a pre-review lottery associated with increased female representation and reduced estimated economic cost. nih + 2

Read carefully: these are surveys and observational studies within specific schemes, not randomised comparisons of funding systems by scientific output. They justify further experimentation; they do not establish that lotteries produce better science.

---

Latest developments

Europe. The Parliament's rapporteur published his draft report on 3 March 2026, with MEPs arguing anything below roughly €200 billion is a real-terms cut; the Council agreed its partial negotiating position on 26 June 2026. Nothing is settled. IdoxERA-LEARN

Talent. Choose Europe for Science now carries close to €900 million — about €400 million more than announced in May 2025 — and ERC Advanced Grant applications from outside Europe nearly quadrupled to 168, with MSCA postdoctoral applications up 65%. Conversion remains modest: 13 of 319 Advanced Grant winners (4%) were based outside Europe, up from 1.4%. European CommissionEuropa

Global. OECD-area government R&D budgets fell 4.1% in real terms in 2024; among nine countries reporting 2025 data, GBARD fell over 5%, including a 7.9% decrease in the United States. China's 15th Five-Year Plan targets at least 7% average annual R&D growth for 2026–2030. OECDThe Defense News

Next year. The FY2027 request seeks $41.2 billion for NIH (about a 13% cut) and roughly $4 billion for NSF (nearly 55%), with cuts also proposed to ARPA-H (-37%), DOE Office of Science (-15%) and NASA Science (-46.3%). Congress diverges: the House committee approved $47.316 billion for NIH including $7.462 billion for NCI, against a community request of at least $51.303 billion. FABBS + 2

---

Common myths

• "Philanthropy will fill the gap." Federal support was $58.6 billion in 2023 against $24.2 billion total philanthropic contribution. SSTI
• "A signed appropriation guarantees funded projects." Obligation timing, terms and termination authority sit downstream.
• "Indirect costs are profit." They are audited institutional costs.
• "Peer review is validated as optimal." Near-universal, but comparative evidence is sparse.

---

Practical takeaways

1. Track three numbers per funder — request, enactment, obligation rate.
2. Build a 12-month bridge policy for postdoc and technician payroll before you need it.
3. Diversify across funder types, not just grants within one agency.
4. Keep F&A documentation audit-ready in case a transparency model advances.
5. Treat termination and international-collaboration clauses as material contract terms.
6. Follow rule-making, not only appropriations.
7. Always state fiscal year, account basis, and request-versus-enacted when citing figures.

---

Key insights

1. The dominant 2026 risk shifted from budget size to administrative discretion.
2. Enacted US budgets came in near flat despite requests for deep cuts.
3. Overhead protection is real but renewed annually via appropriations riders.
4. FAIR is the most developed transparency-oriented alternative in play.
5. Obligation lag is now a legitimate financial-planning variable.
6. Philanthropy operates at roughly one-sixth the scale of federal support in the relevant US segment.
7. Europe's framework budget remains contested between €175bn and €200bn+.
8. Talent schemes show large application growth and small award growth.
9. China's R&D intensity has crossed the OECD average, with growth written into national planning.
10. Allocation reform is empirically live but not yet decisive.

---

FAQ

What are indirect costs? Shared infrastructure costs — buildings, utilities, compliance, data security — that support many projects and cannot be attributed to one.

Did the NIH 15% cap take effect? No. Courts found NIH violated statutory law and its own regulatory procedures, and no Supreme Court petition followed. Higher Ed DiveACS C&EN

What is the FAIR model? A JAG recommendation replacing the F&A structure, differentiating by institution and research type and removing university-wide rate negotiation. Association of American Universities

What was enacted for NIH in FY2026? $47.2 billion base, with ARPA-H at $1.5 billion. American College of Radiology

What is the OMB grants rule? A proposed government-wide change expanding political appointees' role, reducing merit review's weight and easing terminations. American Association for Cancer Research

How much will Europe spend? €175 billion proposed for 2028–2034, still under negotiation. europa

Is EU funding easier for non-Europeans now? Applications rose far faster than awards. 4% of 2025 Advanced Grant winners were based outside Europe. Europa

Can philanthropy replace public funding? No — the Science Philanthropy Alliance concludes it works best as a catalytic complement. Sciphil

Do funding lotteries work? Early evidence is suggestive; no robust comparison of scientific output across allocation systems exists.

What is the return on public research funding? Advocacy estimates cite $2.57 in economic activity per $1 and over 390,000 jobs, but methodologies and counterfactuals vary widely. AAI News

---

Glossary

Appropriation — binding congressional grant of budget authority. Obligation — the agency action legally committing funds. F&A — facilities and administrative (indirect) costs. GBARD — government budget allocations for R&D. GERD — gross domestic expenditure on R&D. CR — continuing resolution extending prior-year funding. FAIR model — proposed alternative F&A framework. FP10 — the EU's 2028–2034 framework programme. Modified lottery — random selection among screened, comparable proposals. Szilard point — where an allocation process consumes more than it distributes.

---

Conclusion

The 2026 environment is best described not as austerity but as volatility with a stable topline. US enacted budgets held roughly flat despite requests for reductions of 40% and more; what changed was reliability downstream of enactment — timing, terms, and the discretion agencies claim over awards already made. In Europe, the topline itself is undecided.

The practical consequence is a shift in what deserves attention: monitoring appropriations is necessary but no longer sufficient. Rule-making, obligation rates and award terms now carry comparable weight, and diversification plus explicit bridge planning are the defensible responses.

The uncertainties should be stated plainly. Whether the proposed grants rule is finalised, revised or withdrawn is unknown at publication; so is the fate of Europe's budget; so, on current evidence, is whether alternative allocation mechanisms improve scientific output. That last gap is the most striking feature of the field: research funding allocates enormous resources using methods rarely subjected to the standard of evidence it demands of everything it funds.

---

SEO appendix

Primary keyword: research funding. Secondary: research grants 2026, NIH budget 2026, indirect costs research, Horizon Europe FP10, science philanthropy, grant peer review. Long-tail: what is the FAIR model for indirect costs; FY2027 NIH budget request explained; do research funding lotteries work. Entities: NIH, NSF, OMB, European Commission, ERC, OECD, Joint Associations Group, Science Philanthropy Alliance. Internal links: Research → Policy; Research → Peer Review Explained; Guides → Writing a Competitive Grant Proposal. External: congress.gov, oecd.org, research-and-innovation.ec.europa.eu, aau.edu, regulations.gov. PAA targets: Who funds scientific research? What are indirect costs? Is NIH funding being cut? Schema: Article + FAQPage + BreadcrumbList. Snippet opportunities: the direct/indirect definition; Table 2; the FAIR answer. Alt text: "Diagram of the three-stage research funding pipeline from budget request to laboratory."

---

References (APA)

Government sources
Congressional Research Service. (2026). NIH funding: FY1996–FY2026 (R43341). https://www.congress.gov/crs-product/R43341
Congressional Research Service. (2026). The National Science Foundation: FY2026 appropriations (R48783). https://www.congress.gov/crs-product/R48783
Office of Management and Budget. (2026). Regulation for federal financial assistance [Proposed rule, OMB-2026-0034]. https://www.regulations.gov/document/OMB-2026-0034-0001
U.S. Court of Appeals for the First Circuit. (2026). Opinion No. 25-1343. https://www.ca1.uscourts.gov/sites/ca1/files/opnfiles/25-1343P-01A.pdf

Official documentation
Association of American Universities. (2025). JAG releases recommendation for updated research funding model. https://www.aau.edu/newsroom/press-releases/joint-associations-group-indirect-costs-releases-recommendations-f-and-a
AAMC. (2026). Tracking NIH awards in FY 2026. https://www.aamc.org/about-us/mission-areas/biomedical-research/publication/tracking-nih-awards-fy-2026
European Commission. (2026). Choose Europe for Science: over 100 initiatives. https://research-and-innovation.ec.europa.eu/news/all-research-and-innovation-news/choose-europe-science-over-100-national-and-regional-initiatives-aim-attract-global-research-talent-2026-01-30_en
OECD. (2026). OECD overall R&D growth stable; government R&D budgets decline. https://www.oecd.org/en/data/insights/statistical-releases/2026/03/oecd-overall-rd-growth-stable-government-rd-budgets-decline-and-reorient-towards-defence.html
Science Philanthropy Alliance & UC San Diego FSIP. (2026). Science Philanthropy Indicators Report. https://indicators.sciphil.org/key-findings

Academic papers
Liu, M., Choy, V., Clarke, P., Barnett, A., Blakely, T., & Pomeroy, L. (2020). The acceptability of using a lottery to allocate research funding. Research Integrity and Peer Review, 5, 3. https://doi.org/10.1186/s41073-019-0089-z
Woods, H. B., & Wilsdon, J. (2024). Funding lotteries for research grant allocation. Research Evaluation. https://doi.org/10.1093/reseval/rvae025

Industry and association reporting
American Association for Cancer Research. (2026, August 11). Cancer Policy Monitor. https://www.aacr.org/professionals/policy-and-advocacy/aacr-cancer-policy-monitor/cancer-policy-monitor-august-11-2026/
Chemical & Engineering News. (2026, April). Trump administration withdraws on fighting for NIH cap on indirect costs. https://cen.acs.org/policy/research-funding/nih-research-funding-indirect-cost-cap-lawsuit/104/web/2026/04
Physics Today. (2026, February 26). Philanthropies selectively mitigate damage from lost federal science funding. https://physicstoday.aip.org/news/philanthropies-selectively-mitigate-damage-from-lost-federal-science-funding

Editorial note: Where figures differ by source — notably NSF's FY2026 total, which varies by account basis — ranges are stated rather than reconciled. Advocacy-derived estimates are flagged as such. The author has no first-hand involvement in any funding programme described.