The FY2027 MURI Kept Its $170 Million and Cut Six Topics. That Arithmetic Changes Who Wins.

Most years, the Multidisciplinary University Research Initiative is a predictable instrument. It has run since 1985. It funds five-year, roughly $7.5 million team awards at U.S. universities. The topic list rotates, the deadlines slide a few weeks, and the research community reads the new topics to learn what the Pentagon's science offices are worried about.

The FY2027 Department of War MURI solicitation, published September 21, 2026, is not quite that. Two things in it are worth reading closely, and neither is in the topic list.

The announcement carries three numbers because three agencies run it jointly: N0001426SF002 (Office of Naval Research), W911NF26S1000 (Army Research Office), and NOFOAFRLAFOSR20260002 (Air Force Office of Scientific Research). Optional white papers are due November 20, 2026 at 5:00 PM Eastern. Technical and eligibility questions close November 6. White paper feedback lands around January 8, 2027, full proposals are due April 9, 2027 at 5:00 PM Eastern, selections are expected in July 2027, and grants start around August 2027.

The topic count fell 27%. The money did not.

The FY2027 NOFO estimates "approximately $170 million dollars pending out-year appropriations" across five years. That is the same figure the FY2026 MURI NOFO used.

FY2026 listed 22 topics. FY2027 lists 16.

Run the division. At the FY2026 ratio, $170 million spread over 22 topics across five years is roughly $1.55 million per topic per year — which, against a stated cap of "up to $1.5M" annually per grant, implies almost exactly one award per topic. That is how MURI has historically felt to applicants: a single-winner competition inside each topic, with a handful of strong teams fighting over one slot.

At FY2027's ratio, $170 million over 16 topics across five years is roughly $2.1 million per topic per year. Against the same $1.5 million cap, that implies something closer to 1.4 awards per topic — or, across the whole solicitation, roughly 22 or 23 awards instead of 22 topics.

The NOFO says this out loud, in language that is easy to skim past: "there is no guarantee that any of the proposals submitted in response to a particular topic will be recommended for funding. On the other hand, more than one proposal may be recommended for funding for a particular topic."

For a university research development office, that sentence is the strategic content of this entire solicitation. Fewer topics with more money behind each one means:

1. Competition per topic goes up in raw numbers — the same national pool of MURI-capable teams now has 16 doors instead of 22.
2. But conditional win probability inside a topic goes up too, because a topic that attracts two excellent, genuinely different approaches can fund both.
3. Second-place proposals become viable. In a strict one-per-topic regime, being the second-best proposal on a topic is worth nothing. In a 1.4-awards-per-topic regime, it is worth something real.

The practical implication reverses a common piece of advice. Teams are often told to avoid topics where a famous group is obviously the presumptive winner. Under FY2027's arithmetic, a strong, methodologically distinct proposal on a crowded topic is a better bet than it was last year — provided it is not a near-duplicate of the presumptive winner's approach. Differentiation, not avoidance, is the play.

The Gold Standard Science clause is new, and it prioritizes institutions

Section 3.A of the FY2027 NOFO is titled "Adherence to Gold Standard Science." It defines GSS, per Executive Order 14303, as science that is reproducible, transparent, communicative of error and uncertainty, collaborative and interdisciplinary, skeptical of its findings and assumptions, structured for falsifiability, subject to unbiased peer review, accepting of negative results as positive outcomes, and free of conflicts of interest.

Then it does two things that do not appear anywhere in the FY2026 MURI NOFO — a document that contains zero references to Gold Standard Science, EO 14303, or EO 14332.

First, it makes submission itself an attestation. "By responding to this NOFO, you explicitly agree" that the institution understands the GSS definition and that "all personnel involved in the proposed project will adhere to all administration policies, procedures, and guidance respecting Gold Standard Science." There is no form to sign and no certification block to complete. The act of submitting is the agreement.

Second — and this is the consequential part — it states that "The DoW will prioritize institutions that have demonstrated success in implementing GSS, in accordance with EO14303 and EO14332."

Read the noun. Not proposals that demonstrate GSS. Institutions.

That is a different unit of competition than MURI has used before. The seven published evaluation criteria in this NOFO are all proposal-level and all familiar: scientific and technical merit; potential to advance fundamental understanding; DoW relevance; PI qualifications and availability; adequacy of facilities and strength of the data management plan; impact of interactions with DoW laboratories and defense-relevant industry; and realism and reasonableness of cost, with the explicit note that cost sharing is not a factor in the evaluation. GSS institutional track record is not among them.

So the prioritization language sits outside the scored criteria — which makes it a selection-stage consideration rather than a review-stage one. Section 3.B adds a companion requirement from EO 14332: proposals "shall include clear benchmarks for measuring success and progress towards relevant goals." That one is actionable in the document you write.

What a research office should actually do with this, before November 20:

• Find out whether your institution has anything to point to. A published research-integrity framework, a reproducibility policy, an open-data or materials-sharing mandate, a conflict-of-interest regime that goes beyond the federal floor. If it exists, the proposal is where it gets cited.
• Make the benchmarks explicit and quantitative. EO 14332 benchmarks are not a hand-wave about "demonstrating feasibility." They are dated milestones with measurable thresholds.
• Treat negative results as a design feature, not an embarrassment. GSS tenet eight is explicit that negative results count as positive outcomes. A work plan with pre-registered falsification tests and a stated decision rule for abandoning a hypothesis is, under this framing, stronger than one that assumes success.
• Do not confuse GSS with cost share. The NOFO is unambiguous that cost sharing carries no evaluation weight. Institutional commitment should be demonstrated through governance and infrastructure, not dollars.

The 16 topics, and who owns them

Topic ownership determines where your white paper goes, which is a mechanical trap discussed below. AFOSR owns Topics 1-5, ARO owns 6-11, and ONR owns 12-16.

AFOSR (1-5): non-linear mixing processes for EUV and X-ray radiation in optically thick systems; tunable goniopolar semiconductors; robustness to novel queries in AI for scientific discovery; unlocking the potential of TEM ptychography; and Science with Optical Cycling Centers for Energy Redistribution (SOCCER).

ARO (6-11): Th-229 enabled by the solid state (TESS); reticular qubit assemblies; design of superomniphobic materials via novel high-throughput multi-scale methods; "Some Like It Hot," on the biogeochemical matrix of fire- and smoke-borne particulates; interfacial superconcentration in extreme multiphase regimes; and atomic and molecular hybrid quantum solids.

ONR (12-16): distinguishing between AI-generated content and human-created content; troposcatter sensitivity to atmospheric and electromagnetic conditions; interface-mediated control of biofilm function; high-entropy electrochemical materials; and bio-inspired autonomous and recurrent self-healing in materials and structures.

The portfolio reads as a quantum-and-materials year with one conspicuous information-operations entry. Topic 12 — telling machine-generated content from human-generated content — is the only topic on the list whose research question has a consumer-facing analogue that is already a multi-billion-dollar commercial problem. Teams bidding it should expect company.

Per-topic resourcing is stated individually in the topic descriptions and generally reads "no more than an average of $1.50M per year for 5 years, supporting no more than 6 faculty investigators." The faculty cap is a real constraint on consortium-building: a six-institution team with one investigator each is permissible; a six-institution team with twelve investigators is not.

The mechanics that disqualify people

White papers do not go through Grants.gov. The NOFO states it flatly: "White Papers must NOT be submitted through the Grants.gov application process." Each agency has its own channel. AFOSR (Topics 1-5) uses a Qualtrics form. ARO (Topics 6-11) accepts email directly to the Research Topic Chief, or physical mail. ONR (Topics 12-16) uses the ONR Submission Portal at submissions.nre.navy.mil, which requires registration — and ONR additionally suggests emailing a copy to the Research Topic Chief with a cc to the MURI coordinator. A team that prepares a white paper and then goes looking for a Grants.gov submission button on November 20 will miss the deadline.

Full proposals do go through Grants.gov, and ONR-topic submitters must include the ONR white paper tracking number they were issued.

Four pages for the white paper, single-spaced, 12-point Times New Roman, 1-inch margins, excluding cover letter, cover page, references, and CV. "White paper pages beyond the 4-page limit may not be evaluated or read." Twenty-five pages for the technical proposal; the budget has no page limit, and the data management plan sits outside the page limit.

Eligibility is narrower than "universities." Only U.S. institutions of higher education with degree-granting science or engineering programs may submit. A University Affiliated Research Center qualifies only if it is part of a U.S. university and not designated an FFRDC. Industry, DoW laboratories, FFRDCs, and foreign entities "may collaborate on the research but may not receive MURI funds directly or via subaward." If an industrial partner needs funding to make the collaboration work, the mechanism is a separate supplemental proposal attached to the primary one, reviewed by the Research Topic Chief, and funded — if at all — from non-MURI or non-Government money. Budget accordingly, and do not plan a subaward line for your industry partner.

Two disqualifying screens worth naming. Per NDAA 2021 Section 1062, since October 1, 2024, DoW may not fund an institution of higher education that hosts a Confucius Institute absent a waiver. And "disclosures of current and pending support made in this application may render an applicant ineligible for funding" — with DoW reserving the right to request updated current-and-pending data throughout the period of performance and to discontinue funding based on it. Research-security hygiene on the PI's disclosures is not a formality in this competition.

What to do in the next six weeks

The white paper is optional and, in the NOFO's own words, "strongly recommended." The reason to submit one is not politeness. It is that the Research Topic Chief will "provide informal feedback notification to the prospective awardees to encourage or discourage submission of proposals," and "may also, on occasion, provide feedback encouraging re-teaming to strengthen a proposal."

Re-teaming advice in early January, four months before a full proposal is due, is the single most valuable output of this process — and it is only available to teams that filed by November 20. A discouraged white paper saves an institution several hundred hours of senior faculty time. An encouraged one with a re-teaming suggestion is a map.

Contact the Research Topic Chief before November 6. The NOFO explicitly recommends communicating with topic chiefs about award size and researcher count before formal submission, and questions submitted inside the final two weeks before a deadline may go unanswered. Six weeks is enough time to file a credible four-page white paper. It is not enough time to assemble a six-investigator consortium from scratch, which is the real argument for starting the conversation this week.