Skip to content

NSF Probability Eligibility Guide: NSF 26-520 (2026)

Updated 7 official sources

Quick Answer: NSF Probability eligibility under NSF 26-520 turns on 7 checks: probability core, subfield fit, application boundary, adjacent DMS route, conference timing, broader impacts, and proposal readiness.

Full answer and how we checked it

Since August 17, 2026, proposals go through NSF 26-520 and are accepted anytime. The old PD 18-1263 target date (third Tuesday in September) no longer applies.

Who this is for: Probability PIs, stochastic-process researchers, applied probability teams, sponsored-research offices, and workshop organizers deciding whether Probability is the right NSF home.

Data note: This guide uses the official NSF Probability page, NSF 26-520, NSF PAPPG guidance, and NSF Awards API records checked on October 12, 2026.

In this guide

Quick Facts

Agency
NSF
Last deadline
Sep 15, 2026 (passed)
Status
Closed

Assistance Listing 47.049: all awards

Source: USASpending.gov FY2024

These figures cover all U.S. National Science Foundation awards under Assistance Listing 47.049, of every award type. They are not figures for the award this guide describes, and one award of that kind can be far smaller.

Total Awards (FY2024, whole listing)
5,546
Avg Award (whole listing)
$1,395,751
Median Award (whole listing)
$375,000
Total Funding (FY2024, whole listing)
$1.6B
YoY Funding Trend
-7.8%

Data last verified: May 19, 2026

According to NSF Probability (opens in a new tab), the program supports the theory and applications of probability, including discrete probability, stochastic processes, random matrices, universality classes, interacting particle systems, stochastic differential equations, and Markov chains and processes. NSF says work applying probability to other areas of science and engineering is especially encouraged. Proposals now go to NSF 26-520 (opens in a new tab), which accepts them anytime.

The decision question is whether probability is the scientific contribution, or whether probability is a tool inside Statistics, Computational Mathematics, Applied Mathematics, CDS&E, data science, operations research, or another NSF route.

Use this guide with the live NSF Probability program record. Then compare NSF Analysis guide, NSF Geometric Analysis guide, NSF Mathematical Sciences Postdoctoral Research Fellowships guide, NSF Mathematical Sciences Infrastructure guide, NSF grants complete guide, how NSF grant review works, what is a NOFO, and the Grantsights grant database.

Source Check

Checked October 12, 2026.

The timing rules changed after the June check. NSF posted NSF 26-520 (opens in a new tab) on August 17, 2026, and the Probability page (opens in a new tab), last updated October 5, 2026, now points to it. Proposals are accepted anytime, so the old PD 18-1263 September target date and its off-window return rule no longer apply.

Source factOctober 12 check
SolicitationNSF 26-520, MPS Mathematical Sciences Research Programs
DirectorateMathematical and Physical Sciences
DeadlineProposals accepted anytime
Old patternThird Tuesday in September (no longer listed)
Conference up to $50,000Submit at least 6 months before the event
Conference over $50,000Submit at least 9 months before the event
Eligible organizationsU.S. two- and four-year IHEs, U.S. nonprofit research or education organizations, federally recognized Tribal Nations
Proposal limitNone per organization or per PI
Cost sharingVoluntary committed cost sharing prohibited

The October 12 source check makes the first-screen answer sharper. This is not just a math grant. The proposal needs a probability contribution strong enough to survive comparison with Statistics, Computational Mathematics, Applied Mathematics, CDS&E, CAREER, MSPRF, and meeting-support routes.

The PI should name the probabilistic object, subfield, proof path, application boundary, and meeting-support need before comparing Statistics, Computational Mathematics, Applied Mathematics, CDS&E, CAREER, MSPRF, and conference or workshop routes.

7 NSF Probability Eligibility Stop Checks

Stop checkStrong Probability signalRed flag
Core questionProbability theory drives the projectProbability is only a model layer
SubfieldFits a listed probability areaStatistics or computation dominates
ApplicationApplication sharpens the probability questionApplication overwhelms the math
Program routeProbability is the best DMS homeAnother program has a clearer panel
TimingDraft is ready and a program director has seen the summaryRushed because there's no deadline
ConferenceMeeting ask is separatedWorkshop support is hidden
EvidenceThe proof path and examples are credibleAims are too broad

Probability proposals can score poorly when they read as application proposals with probability vocabulary added late.

5 Source Facts Applicants Should Know

According to NSF, the Probability Program supports theory and applications of probability. NSF names discrete probability, stochastic processes, random matrices, universality classes, interacting particle systems, stochastic differential equations, and Markov chains and processes.

According to NSF, conference, workshop, and summer or winter school proposals in Probability go through NSF 26-520 as "conference" or "travel" proposals, at least 6 months ahead for requests up to $50,000 and 9 months ahead above that. NSF 26-520 lists no limit on proposals per organization or PI, prohibits voluntary committed cost sharing, and gives no fixed award range; budgets should match the scope of the work.

According to USAspending.gov (opens in a new tab) data, assistance listing 47.049 (Mathematical and Physical Sciences, all MPS programs, not just Probability) had about $1.60B in FY2024 obligations.

FactSource-backed read
SolicitationNSF 26-520 (formerly PD 18-1263)
DirectorateMathematical and Physical Sciences
DivisionMathematical Sciences
DeadlineAccepted anytime under NSF 26-520
Assistance listing47.049

What Probability Has Funded Recently

Grantsights pulled Probability awards from the NSF Awards API (opens in a new tab) on October 12, 2026, covering awards dated October 1, 2025 through that day. There were 22 records, all to universities except one conference award to Vanderbilt University Medical Center: 7 conference or travel awards, 3 CAREER awards, and 12 research awards. Some records are transfers when a PI changed institutions: three tie to one PI's KPZ project moving from Kansas to Maine, and one CAREER award moved to Northwestern. Several of those records also carry EPSCoR co-funding, which is worth reading about in the NSF EPSCoR guide if your state qualifies.

AwardOrganizationTopicEstimated total
2553117New York UniversityRandom planar maps and scaling limits$145,000
2550762UCLANon-mean-field random matrices$145,000
2552511Stanford UniversityTimelike Liouville theory$200,000
2550535University of ChicagoLiouville quantum gravity$300,000
2553445Brown UniversityInteracting particle systems on sparse graphs$489,060
2540987Harvard UniversityCAREER, algorithmic spin glass theory$438,905
2542899University of ChicagoCAREER, geometric foundations of stochastic dynamics$440,000

Grantsights analysis of these records shows a clear size pattern: five research awards were exactly $145,000, every conference award was $50,000 or less, and the three CAREER awards ran $400,000 to $440,000. If you're early career, the NSF CAREER award guide covers that route. Random matrices, KPZ universality, and Liouville quantum gravity came up repeatedly. NSF's current subfield list now names random matrices and universality classes.

What NSF 26-520 Changes

NSF 26-520 (opens in a new tab) covers 11 core math programs, Probability among them, plus focused research groups, infrastructure, and research training groups. NSF estimates 400 to 600 awards and $200,000,000 across the whole solicitation. A common mistake is reading "accepted anytime" as "decided quickly." Proposals still go through merit review, so ask the Probability program directors at PCF@nsf.gov about likely timing. NSF 26-520 also says conference proposals up to $200,000 don't need external review.

What NSF Doesn't Tell You

Within 21 days, answer 4 practical questions. First, probability applications are welcome, but the mathematical contribution must still be probability. A proposal about a biological system, network, or AI method needs to show what probability theory changes.

Second, the route comparison should be written before the abstract. For example, a project proving a new limit theorem for interacting particle systems fits Probability, while a project building predictive models for a domain dataset may need Statistics, Computational Mathematics, or CDS&E.

Third, open submission removes the forcing date, so set an internal one.

Fourth, route triage should test the actual aims, proof maturity, application boundary, and panel fit.

6 Evidence Assets That Make This Proposal Stronger

Within 21 days, build 6 evidence assets. Asset 1 is the probability-core memo. State the probabilistic object, theorem, method, and why the work belongs in probability.

Asset 2 is the subfield map. Match the project to discrete probability, stochastic processes, random matrices, universality classes, interacting particle systems, stochastic differential equations, Markov chains and processes, or related areas.

Asset 3 is the application-boundary memo. If another field motivates the work, explain what remains mathematically new after the application is removed.

Asset 4 is the adjacent-route map. Compare Probability, Statistics, Computational Mathematics, Applied Mathematics, CDS&E, CAREER, MSPRF, and conference support.

Asset 5 is the review-risk table. Include proof risk, modeling risk, computation risk, data-dependence risk, and reviewer-fit risk.

Asset 6 is the broader-impact plan; the NSF broader impacts guide covers what reviewers look for. Include training, seminars, open notes, software when real, or participation plans with enough detail to evaluate.

5 Route Choices Before Drafting

Within 21 days, compare 5 routes.

RouteBetter first read when...Why
ProbabilityProbability theory is the core contributionNSF 26-520 with Probability selected fits
StatisticsInference or statistical method dominatesDMS Statistics may fit
Computational MathematicsAlgorithm design dominatesComputation panel may fit
Applied MathematicsDomain model is centralApplied route may fit
CDS&EData-enabled cross-field work dominatesCross-directorate route may fit

The proposal should not make reviewers infer this comparison. State it clearly.

6 Bad-Fit Signals

First, the application is more important than the probability result. Second, the project is mainly statistical inference. Third, the work depends on a domain dataset more than a theorem.

Fourth, the proposal is really a computational-method project. Fifth, a conference request misses the 6- or 9-month lead time. Sixth, the broader-impact plan is only a generic student sentence.

Those signals do not mean the idea is weak. They mean the route may be wrong.

Warning: if the abstract reads naturally without the probability theorem, don't send it to Probability first. That's usually a sign the project needs Statistics, Applied Mathematics, Computational Mathematics, or a domain program.

21-Day Probability Readiness Plan

Days 1 through 3 should confirm the NSF 26-520 rules, program fit, and whether the ask is research or meeting support. Days 4 through 7 should write the probability-core memo.

Days 8 through 12 should map subfield fit and adjacent DMS routes. Days 13 through 16 should refine proof path, examples, training, and broader impacts. Days 17 through 21 should check PAPPG details, budget logic, project description structure, and review fit.

6 Final Research Checks Before Submission

Within 21 days, finish 6 final checks. Check 1 is the NSF program page. Check 2 is NSF PAPPG (opens in a new tab) compliance. Check 3 is Probability route fit. Check 4 is the NSF 26-520 solicitation. Check 5 is conference or workshop timing. Check 6 is adjacent DMS fit.

You're ready when the proposal can state the probability contribution in 2 sentences and explain why Statistics, Computational Mathematics, Applied Mathematics, or CDS&E is not a better first route.

4 Decision Questions Before Writing

Within 21 days, ask 4 decision questions. Question 1: "Is probability theory the main contribution?" Question 2: "Which listed subfield fits?" Question 3: "Would Statistics or Computational Mathematics review this better?" Question 4: "Has a program director seen the one-page summary?"

If question 3 points elsewhere, compare those routes before the PI writes aims.

3 Last-Mile Checks For Probability PIs

Within 21 days, run 3 last-mile checks. Check 1 is probability-center proof. Check 2 is route proof. Check 3 is solicitation proof.

If those checks are weak, a program officer conversation may save a rewrite later. A final note for applicants: write the fit sentence before the abstract. In 2 sentences, state the probabilistic object, method, contribution, and reason Probability is the best NSF home.

Sources

Largest awards active in FY2024

One row per recipient, its largest award | 5,546 Assistance Listing records | Avg: $1.4M

REAL DATA
1

Association of Universities for Research in Astronomy, Inc.

DC

$569.7M
2

Associated Universities Inc

VA

$466.8M

Use the visible recipient rows as free source context

This program has no reviewed paid report today.

No reviewed paid report is currently offered for this program. The visible source rows remain free.

Next steps

Where to go next

These next pages help you move from research into a specific grant choice.

Frequently Asked Questions

What does NSF Probability support?

It supports theory and applications of probability. NSF lists discrete probability, stochastic processes, random matrices, universality classes, interacting particle systems, stochastic differential equations, and Markov chains and processes.

What program number applies?

NSF 26-520, the MPS Mathematical Sciences Research Programs solicitation, posted August 17, 2026. The program used to run under program description PD 18-1263, which no longer applies.

What is the next target date?

There isn't one. NSF 26-520 accepts Probability proposals anytime, and the September target date is no longer listed.

Can conference proposals go here?

Yes, through NSF 26-520 as a conference or travel proposal with Probability selected. Submit at least 6 months before the event for requests up to $50,000, or 9 months ahead for larger requests. The event must be held at a U.S.-based institution.

Who should skip this route?

Skip it when statistics, computational mathematics, applied mathematics, data science, or another NSF program is the better review home.

Last updated October 12, 2026. Corrected October 12, 2026: what changed. Published by Grantsights from public sources. See our editorial policy, how we build and check guides, corrections and how we make money.