Page plan · 20 pages

Written for: Policy students and readers of data journalism

1. 1

   Federal computing research funding, FY 2024 to FY 2026

   • This editorial reviews federal computing research funding from FY 2024 to FY 2026.
   • Our evidence comes from government budget records, not press releases or commentary.
   • We follow the money from agencies to national labs, then to the risks ahead.
   • The aim is to show what the numbers say and where they stop.
2. 2

   Why budget records tell this story

   • Budget records are public, comparable year to year and published by the agencies themselves.
   • They separate actual, estimated and proposed amounts, which tells readers how firm a number is.
   • Proposals are requests, not enacted spending, so they should be read with caution.
   • Records show money committed, not the results of that money.
3. 3

   From agency budgets to national labs and risks

   • Part one looks at agency budgets and which agencies carry the largest computing portfolios.
   • Part two follows funding into the national laboratories.
   • Part three examines risks to the funding pipeline.
   • The route moves from budget lines to institutions to open questions.
4. 4

   How grants are spread across the fiscal year

   • Grants do not arrive evenly; award timing follows the fiscal year, which runs October to September.
   • Many agencies cluster awards late in the year as budgets are obligated.
   • Timing matters for researchers planning staff and equipment.
   • Annual totals can hide when money actually reaches projects.
5. 5

   Spending on shared computing centres, FY 2024 to FY 2026

   • Funding for high-capability computing infrastructure and applications was US$1,897.4 million in FY 2024 (actual).
   • It was US$1,894.1 million in FY 2025 (estimated), almost unchanged.
   • The FY 2026 proposal is US$1,429.8 million, a clear drop.
   • That is about US$465 million below FY 2024, roughly a quarter less.

   Column chart: NITRD funding for high-capability computing infrastructure and applications (US$ million). Source: Networking and Information Technology R&D Supplement to the President's FY 2026 Budget Request (Tables 1-4) · checked 2026-10-10

6. 6

   Which agencies fund the most computing research

   • In FY 2024, HHS/NIH had the largest NITRD budget at US$2,983.4 million.
   • DOW excluding DARPA followed at US$2,308.6 million, then NSF at US$2,014.7 million.
   • DOE excluding NNSA reported US$1,550.0 million and DOW/DARPA US$1,270.1 million.
   • DOC/NIST was the smallest of the six at US$223.9 million.

   Bar chart: Largest NITRD budgets by agency, FY 2024 actual (US$ million). Source: Networking and Information Technology R&D Supplement to the President's FY 2026 Budget Request (Tables 1-4) · checked 2026-10-10

7. 7

   Funding against published output

   • This slide sets funding beside published research output.
   • Output counts are an imperfect proxy for research value.
   • Money and publications rarely move in step, because results lag spending by years.
   • Any comparison should use consistent definitions of both measures.
8. 8

   Who funds a typical project

   • A typical project often draws on more than one funder.
   • Combining sources spreads risk but adds reporting burdens.
   • Illustrative example: a US$1 million project split across three agencies (illustrative).
   • Budget records by agency rarely show these shared arrangements clearly.
9. 9

   NIH, NSF and DOE across three budget years

   • NIH falls from US$2,983.4 million in FY 2024 to US$2,970.6 million in FY 2025, then US$1,754.9 million proposed.
   • NSF goes from US$2,014.7 million to US$1,961.9 million, then US$836.4 million proposed.
   • DOE rises from US$1,550.0 million to US$1,676.8 million, then US$1,411.1 million proposed.
   • NSF's proposed cut is the steepest of the three; DOE's is the mildest.

   Column chart: NITRD budgets for NIH, NSF and DOE (US$ million). Source: Networking and Information Technology R&D Supplement to the President's FY 2026 Budget Request (Tables 1-4) · checked 2026-10-10

10. 10

    Five research regions ranked

    • This slide ranks five research regions by funding received.
    • Rankings show concentration, with a few places drawing most awards.
    • Rank order says nothing about quality, only about where money lands.
    • Population and institution size shape the ordering.
11. 11

    Two funding models side by side

    • Two funding models can be compared: competitive grants and directed institutional funding.
    • Competitive grants reward proposals reviewed by peers.
    • Directed funding gives stable support to a facility or program.
    • Each model trades flexibility against predictability.
12. 12

    The second tier of funders

    • Beyond the largest agencies sits a second tier of smaller funders.
    • Their budgets are smaller but often target niche needs.
    • Smaller programs can be more exposed to cuts.
    • Together they widen the base of support for computing research.
13. 13

    Part two: the national laboratories

    • Part two turns to the national laboratories.
    • Labs are federally funded institutions that host large shared facilities.
    • They run long-term programs that universities rarely can.
    • Their budgets come mainly through agencies such as DOE.
14. 14

    Where lab money flows

    • Lab money flows from agency appropriations to lab operations and facilities.
    • Part goes to staff, part to computing hardware and upkeep.
    • Some funds pass on to universities and partners through subcontracts.
    • Following these flows is harder than reading top-line budgets.
15. 15

    Few projects, most of the spend

    • Spending is often concentrated in a few large projects.
    • A handful of flagship efforts can take most of a program's budget.
    • Concentration makes totals sensitive to one project's schedule.
    • Illustrative example: five projects taking 70% of a budget (illustrative).
16. 16

    A flagship supercomputer hub

    • A flagship supercomputer hub shows how shared computing centres work.
    • Such hubs serve researchers across many fields and institutions.
    • They need steady funding for hardware, power and staff.
    • Upgrade cycles span years, so budget swings create planning problems.
17. 17

    Part three: risks to the funding pipeline

    • Part three examines risks to the funding pipeline.
    • Proposed cuts, delayed appropriations and shifting priorities all threaten continuity.
    • Risks compound when several hit the same program.
    • The next slides compare spending with results and group the risks.
18. 18

    Spending against results

    • Spending and results are linked, but not in a simple line.
    • Results such as papers, tools and trained students appear years after funding.
    • Short-term cuts may show harm only much later.
    • Better measures would help policymakers judge returns.
19. 19

    Four risks on a grid

    • Four risks are laid out on a grid of likelihood and impact.
    • Budget cuts are the most visible risk in the records.
    • Delays in appropriations disrupt grant timing.
    • Aging infrastructure and talent loss are slower but lasting risks.
20. 20

    What to watch in the next budget cycle

    • Watch whether FY 2026 proposals are enacted or changed by Congress.
    • Compare final appropriations with the proposed figures shown here.
    • Track whether shared computing centre funding recovers or keeps falling.
    • Check next year's supplement for updated actual figures.