NSF Just Scrapped the Four-Proposal Cap on Instrumentation Grants — and Cut the Pot by a Third

For more than a decade, the hardest part of winning a National Science Foundation Major Research Instrumentation award was not convincing NSF. It was convincing your own university. Each campus could send only a handful of proposals, so every research office ran an internal contest months ahead of the federal deadline, and most faculty with an instrument need were eliminated by colleagues before a reviewer ever read a page.

That system ended on November 6, 2026. NSF posted NSF 27-500, the replacement for the long-archived NSF 23-519. Its revision notes are three sentences long, and one of them changes how every research university handles shared equipment: "Organizational submission limits have been eliminated."

The same document cuts the money. NSF 23-519 projected $75 million for about 100 awards. NSF 27-500 projects $50 million for about 60. Open the door wider and shrink the room behind it, and you get the most consequential change to the program since the CHIPS and Science Act waived its cost-sharing requirement in 2022.

What NSF 27-500 Actually Changes

The core terms, taken from the solicitation:

• Submission windows: January 4 – January 29, 2027; October 18 – November 19, 2027; and October 16 – November 17, 2028. Each closes at 5 p.m. local time for the submitting organization.
• Funding: about $50 million and about 60 awards, subject to appropriations and proposal quality.
• Request range: $100,000 to $4 million. Proposals in mathematics, and proposals from non-Ph.D.-granting institutions in any NSF-supported field, may ask for less than $100,000.
• Limits: none per organization, and none per PI or co-PI.
• Tracks: eliminated. Helium-recovery equipment proposals are still reviewed separately from research instruments.
• Duration: up to 36 months for acquisition and up to 60 months for development.
• Earliest start: October 1, 2027, for proposals submitted in the January 2027 window.
• Cost sharing: voluntary committed cost sharing is prohibited.
• Eligibility: two- and four-year institutions of higher education, non-academic nonprofits, and federally recognized Tribal Nations.

Two structural rules carry over. In an acquisition proposal, at least 70 percent of total project cost must be on the Equipment line. Any request over $1.4 million must explain the instrument's impact on its research community at the regional or national level. That $1.4 million line is where the old Track 1 ended and Track 2 began. The track labels are gone, but reviewers still apply the higher bar above it.

The End of the Internal Competition

Under NSF 23-519, each performing organization could submit two Track 1 proposals (under $1.4 million), one Track 2 proposal ($1.4 million to $4 million), and one Track 3 helium-conservation proposal. That was four slots for an entire university system's worth of chemists, physicists, engineers and biologists.

Universities managed that scarcity with limited-submission offices. As we reported in October, internal white papers for the expected fall 2026 window were due in mid-August at institutions such as Boise State and CU Boulder. Those internal contests are now pointless. Any eligible PI can submit.

That sounds like pure good news. It isn't, for three reasons.

Volume will rise sharply. The cap held back an unknown but clearly large amount of demand. MRI also skipped its FY2026 competition entirely, so instrument needs have been building since the November 2024 window. Take away the gate, cut the award count by 40 percent, and the success rate will almost certainly drop below what MRI applicants are used to. NSF has not published a projected funding rate. Budget your time as though this is a long-shot competition.

Your colleagues are now your competitors in front of NSF, not before it. Two groups at one university proposing overlapping mass spectrometers used to be sorted out internally. Now both proposals can reach the same panel, and the panel will notice. Each proposal has to show that the instrument is not available elsewhere on campus. Two near-identical requests from one institution weaken each other.

The institution still has to sign. Every proposal needs a one-to-two-page operations commitment letter from the institution that lists its MRI awards from the previous five calendar years and commits to operating and maintaining the instrument. That letter makes the provost's or research office's endorsement a real gate even without a formal cap. Research offices that ran limited-submission reviews should not disband them. They should turn them into a coordination process that keeps duplicate proposals out, makes sure operations funding exists, and catches budget problems early.

Less Money, Fewer Awards, Bigger Average

Divide the new numbers and the average award rises, from about $750,000 under 23-519 ($75 million across 100 awards) to about $833,000 under 27-500 ($50 million across 60). That is a projection, not a policy. Still, it suggests NSF expects to fund fewer, larger instruments rather than spreading the money thin.

The solicitation does not explain the cut. The program's own history gives some context: NSF used FY2026 money to fund meritorious leftovers from the FY2025 competition instead of running a new one. A $50 million projection may reflect an agency that has already committed part of its instrumentation budget, or one planning around FY2027 appropriations that are still unsettled. In either case, the practical point is the same. Sixty awards nationwide, across every NSF directorate, with no institutional throttle, will be very competitive.

For primarily undergraduate institutions and community colleges, this has one advantage. Their sub-$100,000 lane survives, and those requests compete in a part of the program that research universities cannot enter. A $90,000 instrument proposal from a non-Ph.D.-granting institution, with a clear research-training plan and a multi-user access plan, is not competing against $3.8 million cryo-EM requests.

The Window That Disappeared

In October, the governing document for MRI was NSF 23-519. Its submission table listed a window of October 15 – November 16, 2026. NSF 23-519 now states that it is archived and that "the latest version is NSF 27-500." The current MRI program page lists only the 2027 and 2028 windows. Neither page says outright what happens to a proposal submitted under the old fall 2026 window, which on paper is still open until November 16.

Read the documents together and the conclusion is clear: plan against NSF 27-500 and the January 2027 window. If your institution submitted, or was about to submit, under 23-519 this fall, email mri@nsf.gov now and get the program's answer in writing before November 16. Don't assume a proposal submitted against archived rules will be reviewed. Don't assume it will be returned either.

The Cost-Share Clock Is Still Running

NSF 27-500 says nothing about mandatory cost sharing. That gap matters.

Mandatory MRI cost sharing comes from statute: section 7036(c) of the America COMPETES Act. Before the waiver, Ph.D.-granting institutions owed 30 percent of total project cost. Section 10320 of the CHIPS and Science Act waived the requirement for five years, starting with the FY2023 competition. NSF 23-519 stated that waiver explicitly. NSF 27-500 does not mention it.

The January 2027 window falls in fiscal year 2027, the fifth year of that waiver. The October 2027 and October 2028 windows fall in FY2028 and FY2029, after a five-year waiver would have run out, unless Congress extends it. That is our reading of the statute, not something NSF has stated. If it holds, a $4 million request in January 2027 needs no institutional match. The same instrument in October 2027 could require a Ph.D.-granting institution to put up about $1.71 million, since NSF's $4 million would then be only 70 percent of a $5.71 million project.

That makes the January window financially distinct from the two that follow. A large-instrument proposal that is close to ready should go in January.

The New Policy Language

Three passages in 27-500 reflect the current administration's priorities.

First, the Priorities section says NSF "expects the highest standards of scientific rigor, integrity, and adherence to tenets of Gold Standard Science." That phrase appears across this year's NSF solicitations. In practice, it means the project description should spell out validation, calibration and reproducibility plans for the instrument, not leave them implied.

Second, the program "especially welcomes" instruments that advance national R&D priority areas, including "scientifically justified integration with SI-enabled workflows." The solicitation does not define "SI." It links to the White House's FY2028 R&D priorities memorandum. If automated, data-intensive or AI-assisted pipelines are part of how your instrument will be used, say so and justify it scientifically. Don't add the language just to match the solicitation.

Third, the solicitation includes the Executive Order 14332 terms now standard across federal grants: NSF "may give preference to institutions with lower indirect cost rates," and awards "may include termination for convenience provisions." On an equipment-heavy MRI budget, indirect costs are a small share, because equipment is excluded from the modified total direct cost base. Even so, institutions with high negotiated rates should watch how that preference is applied.

What to Do Before January 4

• Re-poll your faculty. People who lost the August internal competition, or who never entered it, can apply now. Tell them by mid-November. A strong MRI proposal needs vendor quotes, user letters and an operations plan, and the window opens in eight weeks.
• Run a de-duplication review, not a selection review. Map every planned proposal against your existing instruments and against each other. Combine overlapping requests into one stronger multi-user proposal.
• Prioritize large instruments for January. If the cost-share waiver ends with FY2027, the January window is the last match-free opportunity for big-ticket requests.
• Write to the review criteria. Panels will score multi-user, shared-use research explicitly. Name the users, departments, partner institutions and expected hours, not just the PI's lab.
• Get quotes now. Itemized vendor quotes are required. Year-end vendor schedules and the holidays make December a bad month to start asking for them.

The four-proposal cap shaped how universities planned around MRI for years, and NSF 27-500 removes it. PIs who move first can use the Granted grant-writing workspace to draft the project description and management plan before the January window fills up.