Skip to content

NSF GRFP Guide 2026: Win a Graduate Fellowship

Updated 7 official sources

Quick Answer: The NSF GRFP pays a $37,000 stipend plus a $16,000 cost-of-education allowance for each of three years, $159,000 per fellowship, and NSF 26-526 estimates about 1,500 new awards this cycle.

Full answer and how we checked it

Applications are due by 8 p.m. ET between October 19 and 23, 2026, depending on field, and reference letters are due earlier, on Friday, October 16. You must be a U.S. citizen, national, or permanent resident and a senior undergraduate, a bachelor's degree holder with no graduate enrollment, or a first-year graduate student; second-year graduate students are not eligible.

Who this is for: Undergraduate seniors planning graduate school, first-year graduate students in STEM or social science fields, and faculty mentors helping students prepare fellowship applications. If you're a U.S. citizen or permanent resident headed into or early in a research-based graduate program, the NSF graduate research fellowship is the single most valuable individual award you can pursue at this stage.

Data note: Award amounts and success rates are based on the most recent federal fiscal year data (FY2024-2025) from agency databases and USASpending.gov. Program parameters may change with annual appropriations.

Source review by Grantsights, June 3, 2026: We checked the current NSF GRFP program page, Research.gov, NSF solicitation materials, NSF FY2026 budget request tables, NSF award data, merit-review criteria, and DDIG-BIO sources against this guide. Where official sources stop: NSF pages state the stipend, allowance, and eligibility rules, but they do not tell applicants how 2025 award compression changes the writing strategy. Based on our analysis, a bad fit is a 2026 applicant reusing pre-2025 advice that treats broader impacts as a small closing paragraph. For example, a life-sciences applicant should use the research plan to show pilot data, named methods, and a specific broader-impact route rather than relying on generic outreach language.

In this guide

The GRFP isn't just money. It's a signal. Departments treat GRFP fellows differently: you'll have more flexibility choosing advisors, more freedom to switch research directions, and stronger positioning for postdocs and faculty jobs later. The GRFP is one of many programs in the NSF grants complete guide, which covers the full range of research, education, and SBIR funding NSF offers. The fellowship follows you if you transfer programs, and that portability changes the power dynamics between you and your department.

Students in EPSCoR jurisdictions should also review the live NSF EPSCoR Graduate Fellowship Program page to see how that state-linked path differs from the national GRFP.

But the application process confuses a lot of people. The two-page personal statement and two-page research plan sound simple until you realize that reviewers spend about 20 minutes on your entire file and that the scoring rubric rewards things most applicants ignore.

This guide breaks down everything: who's eligible, what the application requires, how reviewers actually score it, and the mistakes that sink otherwise strong candidates.

From Manusights, our sister product

Writing an NSF proposal? Manusights offers an AI critique of your proposal against the review criteria in your solicitation: $149 for a research proposal, $49 for a GRFP or other fellowship application. Back within 3 business days of your upload.

Order a grant review

How the NSF Graduate Research Fellowship Works

The GRFP is NSF's oldest fellowship program, running since 1952. It's designed to identify and support outstanding graduate students in STEM and the social sciences early in their careers. NSF explicitly says they're investing in people, not projects.

Here's the financial breakdown:

ComponentAnnual AmountTotal (3 Years)
Stipend$37,000$111,000
Cost-of-education allowance$16,000$48,000
Total fellowship value$53,000$159,000

The stipend goes directly to you. The cost-of-education allowance goes to your institution to cover tuition and fees. If your tuition exceeds $16,000 (which it will at most private universities), the department or graduate school typically covers the gap. At public universities, the allowance often covers the full cost.

You receive three years of support spread over a five-year fellowship period. That means you can "reserve" a year of funding, useful if your advisor has grant support for you during some years but not others.

GRFP Eligibility: Who Can Apply

Eligibility rules trip up many applicants. Here's what you need:

Citizenship: U.S. citizens, U.S. nationals, and permanent residents only. DACA recipients and visa holders aren't eligible. This is a statutory requirement that NSF can't waive.

Degree status: Under NSF 26-526 (opens in a new tab), you must be one of these:

  • An undergraduate senior, or a bachelor's degree holder with no prior enrollment in a graduate degree program
  • A joint bachelor's-master's student in the final undergraduate or first graduate year of the joint program
  • A first-year graduate student in your first graduate degree program, with less than one academic year completed

Second-year graduate students are not eligible to apply.

Application limits: Seniors and bachelor's degree holders can apply an unlimited number of times before enrolling in a graduate degree program. Joint bachelor's-master's students can apply once, in the final undergraduate or first graduate year. Graduate students can apply once, in the first year of their first graduate program. Everyone is limited to one application per annual competition.

Field of study: Your proposed research must be in an eligible field listed in the solicitation's appendix. NSF 26-526 sets deadlines for ten field groups: Life Sciences; Computer and Information Science and Engineering; Materials Research; Psychology; Social Sciences; Engineering; Chemistry; Geosciences; Mathematical Sciences; and Physics and Astronomy. Professional and mostly course-based programs, including business, public health, social work, law, medicine, and clinical psychology, are ineligible, and so is research whose goals are directly disease- or health-related.

Joint degree programs: Joint or combined professional degree-science programs, such as MD/PhD or JD/PhD, are ineligible under NSF 26-526, as are dual professional degree-science programs.


Get research funding calls by email

About once a week, we email open research funding calls with a real deadline, from federal agencies such as NIH and NSF, picked for research institutions. One click to unsubscribe.

No spam. Unsubscribe anytime.

NSF GRFP Application Components

The application has three parts you write and one part others write for you. Everything goes through NSF's Research.gov portal.

Personal Statement (3 pages)

This isn't a personal essay. NSF wants you to demonstrate your preparation for graduate research and your potential for broader impacts. The solicitation calls it the "Personal, Relevant Background and Future Goals Statement."

What reviewers actually look for:

  • Research experiences you've had and what you learned from them
  • How those experiences prepared you for your proposed graduate research
  • Your plan for contributing to broader impacts during and after the fellowship
  • Evidence of leadership, teaching, mentoring, or outreach activities

The biggest mistake: writing a memoir. Reviewers don't need your life story. They need evidence that you can do research and that you've thought about how your work connects to society. Every paragraph should answer the implicit question: "Why does this make you a stronger researcher?"

Research Plan (2 pages)

This is a mini-proposal describing a research project you'd like to pursue. NSF says it doesn't have to be your actual dissertation project, but it should demonstrate your ability to frame a research question, propose methods, and think about significance.

Structure that works:

  1. Introduction and motivation (1/3 page): What's the problem and why does it matter?
  2. Background (1/3 page): What's been done and what's the gap?
  3. Proposed research (2/3 page): What will you do and how?
  4. Broader impacts (1/3 page): How does this work benefit society beyond the lab?
  5. Timeline (brief): When will you do each part?

Two pages forces brutal concision. Don't waste space on textbook-level background. Assume your reviewer is a scientist but not necessarily in your subfield.

Reference Letters (3 required)

You choose three recommenders who submit letters through Research.gov. At least one should speak to your research abilities, and at least one should speak to your broader impacts potential. Faculty advisors from research experiences are strongest. Letters from employers or course instructors carry less weight unless they directly observed your research work.

Recommenders need time. Give them at least four weeks and send them your personal statement and research plan drafts so they can write specific, aligned letters.


NSF GRFP Review Criteria: Intellectual Merit and Broader Impacts

Every NSF program uses two review criteria, and the GRFP is no exception. Reviewers score your application on both:

Intellectual Merit

This criterion asks: does the applicant have the potential to advance knowledge? Reviewers look at:

  • Quality of your research experiences
  • Strength of your academic record
  • Thoughtfulness of your research plan
  • Whether you understand the state of your field
  • Whether your proposed work could produce meaningful results

You don't need to have published papers, but you need to show that you've done real research work, not just followed protocols someone handed you.

Broader Impacts

This criterion asks: how will the applicant benefit society? Reviewers look at:

  • Plans for teaching, mentoring, or outreach
  • How your research connects to societal needs
  • Activities that broaden participation in STEM (especially from underrepresented groups)
  • Plans for science communication or public engagement

This is where most applications fail. Many STEM students treat broader impacts as an afterthought, tacking on a vague sentence about "mentoring undergraduates" at the end. The strongest applications weave broader impacts throughout both documents. They describe specific plans: "I'll develop and lead a workshop series at [specific program] that introduces crystallography to high school students from Title I schools in [city]."

How Scoring Works

Each reviewer assigns ratings of Excellent, Very Good, Good, Fair, or Poor for both intellectual merit and broader impacts. They then write short narrative comments explaining their ratings. A panel of reviewers discusses the top-rated applications and makes funding recommendations.

In practice, you need Excellent ratings on both criteria from most reviewers to be funded. A "Very Good" in broader impacts with "Excellent" in intellectual merit often falls below the funding line. That's how seriously NSF takes it.


How Many GRFP Fellowships NSF Offers

NSF 26-526 estimates about 1,500 new fellowships in this competition, pending available funds. That's the solicitation's planning number, not a guarantee.

Award-offer listOffered-award recordsSource
20251,500Research.gov GRFP award-offer list, checked June 18, 2026
20262,597Research.gov GRFP award-offer list, checked June 18, 2026
This cycle (NSF 26-526)About 1,500 estimatedNSF 26-526 program page

NSF doesn't publish GRFP application counts or success rates by field in the solicitation, and Grantsights doesn't estimate them. To see how many awards went to your field and university, read the award-offer list itself; the GRFP winners and award history guide explains what it shows.


GRFP Application Timeline and Deadlines

Under NSF 26-526, applications must be received by 8:00 p.m. ET on your field's deadline, and reference letters are due before any application:

Date (2026)What is due
Friday, October 16, 8 p.m. ETReference letters, through the NSF Reference Letter System in Research.gov
Monday, October 19Life Sciences applications
Tuesday, October 20Computer and Information Science and Engineering; Materials Research; Psychology; Social Sciences
Thursday, October 22Engineering
Friday, October 23Chemistry; Geosciences; Mathematical Sciences; Physics and Astronomy

A common mistake is treating the reference letters as an afterthought. They're due October 16, before every field's deadline, and NSF won't accept them after 8 p.m. ET that day. You need at least three letter writers.

Research.gov has been unreliable near deadlines in past years. Submit at least 48 hours early. The system allows you to revise your application up until the deadline, so submitting early doesn't lock you in.


What the Smaller Award Count Means for Your Application

The 2025 award-offer list held 1,500 offered awards, and NSF 26-526 plans for about 1,500 again, fewer than the 2,597 offered in 2026. If your advisor's sense of the odds comes from a larger year, adjust it.

Applications that get funded in a tighter year tend to have one thing the rejected ones don't: a specific research plan. Not "I plan to study X using Y method" but "I will test the hypothesis that [specific claim] using [specific dataset], and if the null hypothesis holds, the implication is [specific consequence]." Reviewers can tell the difference between a student who has started the work and one who has only thought about it.

What the NSF GRFP Website Doesn't Tell You

Here's what you won't find in the solicitation.

Review panels vary wildly in caliber and consistency. Your application might be read by a senior professor who's reviewed 50 GRFPs or by a postdoc doing their first review. Panel composition changes every year, and the quality of feedback varies. This means there's inherent randomness in outcomes. If you don't get funded, the problem might not be your application. Successful applicants who reapply after a rejection often get funded the second time with only minor revisions.

Program officers read the panel summaries, but they rarely override them. In theory, a program officer can override panel recommendations. In practice, for the GRFP, this almost never happens because the volume is too high. Your fate is effectively decided by the review panel.

"Honorable Mention" is more useful than you think. NSF names Honorable Mentions alongside its award offers on the same Research.gov list. It doesn't come with money, but it's a recognized distinction you can put on your CV, and many graduate programs provide supplemental funding to students who received Honorable Mention. Some departments treat it as equivalent to having external fellowship funding for the purposes of admission decisions.

The broader impacts revolution is real. Ten years ago, you could get away with a mediocre broader impacts statement if your intellectual merit was strong. That's no longer true. NSF leadership has pushed hard for equal weighting of both criteria, and panels now take it seriously. An application with Excellent-rated intellectual merit can still be rejected when broader impacts is rated Good.


How to Strengthen Your NSF Graduate Research Fellowship Application

Beyond avoiding common mistakes, here are strategies that distinguish funded applications:

Study successful applications. Many past GRFP fellows post their winning personal statements and research plans online. Alex Lang maintains a widely-used collection of successful essays (opens in a new tab). Read 10 to 15 of these in your field before you start writing.

Get feedback from multiple sources. Your research advisor should review your research plan. A faculty mentor or writing center should review your personal statement. If your university has a fellowships office, use it. They've seen hundreds of applications and know what works. Get at least three rounds of feedback before submitting.

Be specific about broader impacts. Don't say "I will mentor underrepresented students." Say "I will mentor two students per year through the [specific program name] at [university], a program that pairs graduate fellows with first-generation college students interested in STEM research. I'll design a six-week introduction to [method] that students complete alongside me in the lab."

Quantify your research experience. Instead of "I worked in Dr. Smith's lab for two years," write "Over 24 months in the Smith Lab, I processed 340 samples using XRF spectroscopy, identified a contamination artifact in the standard protocol, and co-authored a methods paper (in review at Journal of Analytical Chemistry)."

Connect your trajectory. The strongest applications tell a coherent story: past experiences led to current interests, which motivate the proposed research, which connects to a broader career vision that includes both scholarship and societal impact. Reviewers should finish your application understanding not just what you've done, but where you're going and why it matters.

For more on writing strong federal grant applications, see our guide on how to write a grant proposal.


Institutions That Produce the Most GRFP Fellows

NSF publishes annual data on GRFP awards by institution. The pattern is consistent: large research universities with active undergraduate research programs and strong fellowships advising offices produce the most fellows. Recent top producers include UC Berkeley, MIT, Stanford, the University of Michigan, Cornell, Georgia Tech, and the University of Washington.

But here's what the data doesn't show: per-capita success rates. A large university that submits 500 applications and gets 40 awards has an 8% rate. A small liberal arts college that submits 10 applications and gets 3 awards has a 30% rate. Many small colleges and HBCUs have strong track records per applicant because their faculty invest heavily in fellowship preparation.

If your institution doesn't have a strong fellowships advising infrastructure, you can still compete. The quality of your research experience and the clarity of your writing matter more than your institution's name. NSF actively works to fund students from diverse institutional backgrounds, and review is blind to institution name.


Planning Your GRFP Around Other Funding Sources

The GRFP can be held alongside other awards, with some restrictions. You can't simultaneously receive another federal fellowship (like an NDSEG or Hertz) while using GRFP funding in the same year, but you can accept one and reserve the other. Many students accept the GRFP and reserve years strategically, using advisor funding or other fellowships in alternate years to stretch their total graduate funding.

If you're also considering the NSF CAREER award track (for faculty positions later), having a GRFP on your record strengthens future applications. NSF values applicants who've demonstrated commitment to the agency's mission early in their careers.

For students interested in research at national labs, the GRFP can fund work at DOE, NASA, or other facilities as long as your graduate institution approves. This flexibility is underused.



What USASpending.gov Shows for "NSF graduate research fellowship" Awards

A search of FY2025 USASpending.gov (opens in a new tab) award descriptions containing "NSF graduate research fellowship" found 173 awards totaling $466.0M.

MetricAmount
Total awards173
Total funding$465,988,314
Average award$2,693,574
Median award$636,000

Top Recipients

RecipientTotal FundingAwards
REGENTS OF THE UNIVERSITY OF MICHIGAN$41,100,7531
TRUSTEES OF THE UNIVERSITY OF PENNSYLVANIA, THE$29,418,8331
NORTHWESTERN UNIVERSITY$28,092,6721
THE TRUSTEES OF PRINCETON UNIVERSITY$18,789,1591
REGENTS OF THE UNIVERSITY OF MINNESOTA$15,931,0771
THE TRUSTEES OF COLUMBIA UNIVERSITY IN THE CITY OF NEW YORK$15,030,8481
THE REGENTS OF THE UNIVERSITY OF COLORADO$14,230,2642
UNIVERSITY OF CALIFORNIA IRVINE$10,664,2831
UNIVERSITY OF CALIFORNIA SANTA CRUZ$10,512,2862
UNIVERSITY OF MARYLAND, COLLEGE PARK$10,371,0391

Top States

StateFunding
California$99,108,244
Massachusetts$50,423,773
New York$34,890,725
Illinois$26,848,145
Pennsylvania$24,454,560
North Carolina$17,531,899
Michigan$16,575,193
New Jersey$14,358,499
Georgia$13,829,160
Washington$12,856,250

Data source: USASpending.gov FY2025 award description search, accessed 2026-04-04.


Sources

Award data for this program

Frequently Asked Questions

What is the NSF GRFP stipend amount?

The NSF GRFP provides an annual stipend of $37,000 for three years of support, used over a five-year fellowship period. In addition, your institution receives a $16,000 cost-of-education allowance per year, which covers tuition and fees. The total award value over three years is $159,000. Your institution cannot charge you tuition above the cost-of-education allowance while you're on the fellowship, and many departments waive the difference entirely.

Can I apply to the GRFP if I'm already in graduate school?

Only in your first year. Under NSF 26-526, first-year graduate students in their first graduate degree program, with less than one academic year completed, can apply once. Second-year graduate students are not eligible. Joint bachelor's-master's students can apply once, in the final undergraduate or first graduate year. Undergraduate seniors and bachelor's degree holders who haven't enrolled in a graduate degree program can apply an unlimited number of times before they enroll.

How competitive is the GRFP?

NSF's current solicitation, NSF 26-526, estimates about 1,500 new fellowships, pending available funds. The Research.gov award-offer lists showed 1,500 offered awards for 2025 and 2,597 for 2026 when Grantsights checked them on June 18, 2026. NSF doesn't publish an application count or success rate by field in the solicitation, so use the award-offer lists to see how many awards went to your field and university.

Do I need a specific research advisor or lab before I apply?

No. NSF funds the person, not the project. You don't need a confirmed advisor, an accepted PhD offer, or even a specific university when you apply as an undergraduate. Your research plan should describe a project you're genuinely interested in, but NSF understands it may change after you start your program. Reviewers care more about your ability to articulate research questions and broader impacts than about the specific project details. That said, demonstrating awareness of existing work in your area strengthens your application.

What's the difference between the GRFP and the now-discontinued DDIG?

The Doctoral Dissertation Improvement Grant (DDIG) was a separate NSF program that funded specific dissertation research costs like fieldwork, lab supplies, and data collection. It didn't provide stipend support. NSF discontinued the DDIG in most directorates starting in 2017, though the Directorate for Biological Sciences brought it back as the DDIG-BIO. The GRFP, by contrast, funds your stipend and tuition for three years. If you need both living support and research costs, the GRFP is the primary vehicle now, and you'd supplement research expenses through your advisor's grants or other sources.

Last updated October 8, 2026. Corrected October 8, 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.