Hertz Fellowship Acceptance Rate: The Real Funnel

Hertz Fellowship Acceptance Rate: The Real Funnel

The Hertz Foundation does not publish an acceptance rate. What it publishes is a funnel, and for the cycle that produced the 2026 class of Hertz Fellows the Foundation states the whole chain on its own site: "2025 saw the highest number of applications the Foundation has ever received: 1493," of which "119 were interviewed," "54 finalists received a second interview," and "19 new 2026 Fellows were chosen."

Divide the ends of that chain and you get about 1.27%. That number is our arithmetic, not the Foundation's — it appears nowhere on hertzfoundation.org, it describes one completed cycle, and it is not a forecast for 2027. Everything below is dated, sourced to the Foundation's own pages, and labelled where it is derived.

Quick answer

• Published funnel (one past cycle, retrieved August 19, 2026): 1,493 applications → 119 interviewed → 54 second interviews → 19 fellows.
• ~1.27% is CALCULATED by us as 19 ÷ 1,493 from two Foundation announcements about that same cycle. It is not a Foundation-published selection rate and not a 2027 projection.
• The application volume was a record, in the Foundation's own words — so this particular denominator is the largest one the program has had.
• The funnel has two interview rounds. The written application does not win the fellowship. Its job is to get read past preselection.
• Comparators are not comparable. NDSEG publishes cumulative totals since inception; NSF publishes GRFP awardee lists per competition year. Neither is a current-cycle acceptance rate, and neither is Hertz.

The funnel the Foundation actually publishes

Retrieved from hertzfoundation.org on August 19, 2026. Two Foundation announcements cover this cycle: the fellows announcement, which carries all four counts, and the February 5, 2026 finalists announcement, which describes the 54 as "Selected from more than 1,450 applicants."

The derived ratios, labelled as derived

These are CALCULATED by GradPilot on August 19, 2026 from the four counts above. The Foundation publishes the counts; it does not publish any of these percentages.

Three cautions attach to every figure in that table, and they matter more than the figures do:

1. It is one cycle. A record application year has an unusually large denominator by definition. A single year's ratio is not a program constant.
2. It is not a 2027 number. The 2027 competition had not closed when this was written; the Foundation's published timeline puts preselection in November 2026 and recipient notification in April 2027.
3. A base rate describes a pool, not a person. It tells you how large the field is. It does not tell you anything about a specific application, and nothing in this post — or in any review, ours included — changes a base rate.

The number that actually informs the writing: 119

The single most useful figure in that chain is not 19 and not 1.27%. It is 119 — because that is the count of applications that got a conversation.

The published sequence is written application → preselection → first-round interviews in December and January → second-round in-person interviews in February. The Foundation's FAQ states it plainly: "First round interviews will take place in December and January. Second round interviews (in-person) for select finalists will be conducted in February 2027."

Read the structure back to front and the written application's role gets very specific:

• The essays do not select the fellows. Two rounds of interviews sit between the written file and the award.
• The essays decide who is in the room. Preselection is a reading stage. Roughly ninety-two percent of that cycle's applicants — 1,374 of 1,493, our subtraction — never reached an interview at all, which means their written file was the entire application.
• So the written application has exactly one job at stage one: survive the read. Not to be complete. Not to be impressive across every dimension. To be legible enough, and specific enough, that a reader moves it forward.

Why the essays and the interviews are testing the same thing

This is the part that makes the funnel useful rather than merely intimidating. The attributes the Foundation asks its readers to look for are not different from the attributes a conversation would surface — they are the same attributes, asked twice in two formats.

The Foundation's Who Can Apply page organizes what it is looking for under five headings: Deep, Integrated Knowledge ("Grades and test scores only tell part of the story. We're looking for applicants with deep scientific and technical knowledge..."), Creative Problem Solvers, Commitment to the Greater Good, Active Leadership ("They require self-starters who can gain the trust of their colleagues, build consensus around an ambitious vision..."), and Lifelong Achievers.

The recommender form asks four other people to assess the same register. Per the Foundation's recommenders page, recommenders assess "degree of fundamental knowledge in their field, self-reliance and independence in scientific work, possession of fertile imagination, and originality in their work," and are asked to "briefly discuss meaningful examples of the applicant's creativity, initiative and problem-solving abilities." The Foundation also explains why it uses a form at all: "We learn more about our applicants through our specific form questions rather than a standard letter."

And the essays are pointed at the same target. There is no research project proposal in this application — "The fellowship is intended to fund individual students, not a specific project. There is no requirement to submit a proposed research project as part of the application process." The suggested lengths are suggestions, not caps: "you may go under or over the suggested length of 250–500 words for the various short essays and 500–1000 words for the Personal Essay."

Put those three surfaces side by side and one working conclusion follows, which is the honest reason a funnel with two interview rounds should change how you write:

A written claim you could not sustain in an ordinary conversation about your own work is the failure mode. Not a weak sentence — an unsupported one. If an essay asserts independence, originality, or creative problem-solving, the same assertion is going to be examined by recommenders who were there and, for applicants who advance, by people asking about it directly.

That is a statement about what to write, and it is where our involvement stops. We review writing. We do not coach interviews, we do not simulate them, and nothing here should be read as preparation for one — the Foundation runs its own process and its published materials govern.

The practical version is a self-check on your own draft, not a technique:

• Take each claim in your essays about how you think or work.
• Ask what you actually did that produced the claim — a decision, a constraint you hit, a thing you built or abandoned, a result you had to explain.
• If the answer is a sentence you would say out loud without hesitating, the claim is anchored. If the answer is a restatement of the claim, the sentence is decoration and it is spending words that the reader at preselection has to get through.

Our general method for reasoning from what an award says it wants back to the evidence you choose is in how to write a scholarship essay that fits the award.

Honest comparison: the other two numbers people cite

Applicants almost always look up Hertz alongside NDSEG and the NSF GRFP, then compare three numbers that are not the same kind of number.

NDSEG publishes cumulative totals, not a rate. The program's home page states: "Since program inception, over 4,700 fellowships have been awarded and over 70,000 applications received." Those are totals accumulated across decades of cycles — not a current-cycle acceptance rate, and dividing one by the other yields a statistic that describes no year you could apply to. We work through why that number resists conversion in the acceptance-rate section of the NDSEG FY2027 essays guide.

NSF publishes GRFP outcomes per competition year, but not a rate either. NSF maintains a public, per-year record of Awards Offered and Honorable Mentions on Research.gov, searchable by award year — so the number of offers in a given competition is a matter of public record you can look up at the source. The solicitation itself commits to no number: NSF 25-547 says only that "NSF will support new Graduate Research Fellowships under this program solicitation pending availability of funds." As of August 19, 2026, NSF 25-547 was still the listed GRFP solicitation; see the FY2027 GRFP cycle guide for what that means for timing.

Three sponsors, three disclosure habits, zero published acceptance rates. Any site quoting you a clean current-year percentage for any of the three has computed it — the question to ask is from what, and when.

What a ~1% ratio should and should not change about October

Hertz, NDSEG, and GRFP all land in the same autumn stretch, and a low ratio is exactly the fact most likely to be misused when you plan that month.

It should not change whether you apply. Eligibility is the decision; the denominator is not. And the Foundation's eligibility window is narrow: applicants must be "college seniors, first-year graduate students, or individuals in a gap year preparing to apply to graduate school." Its FAQ adds that the Foundation "generally do[es] not award fellowships to students after their first year of graduate school, except in rare cases where the fellowship will make an exceptional difference in the kind, quality, or creativity of the applicant's graduate research." A ratio does not extend that window, and it does not come back next year.

It should not turn into padding. The most common response to a scary number is to add — more claims, more breadth, more adjectives. A reading stage that moves about 8% of files forward is not rewarding volume; it is rewarding a file that is easy to read and hard to doubt.

It should change how you allocate the month. Three fall applications do not share a contract — GRFP requires named Intellectual Merit and Broader Impacts headings, NDSEG requires redaction and treats a leak as grounds to stop an application advancing, and Hertz requires no research proposal at all. Budget by contract, not by word count. The GRFP, NDSEG and Hertz portfolio guide sets out the deadline ladder and what evidence transfers between the three.

It should push the anchoring work earlier. Because the written file is the whole application for most applicants, the sentence that cannot survive a follow-up question is more expensive here than in a program where a committee reads once and decides. Anchoring claims to specific decisions you made is October work, not November work.

And it should be held loosely. One cycle, a record denominator, no published rate, no 2027 number. Re-read the Foundation's own pages when the application opens.

Where a review fits

The Hertz application and its published terms govern everything about what you submit — including what help is permitted. The terms you agree to inside the 2027 application — including anything the Foundation says about how your application must be prepared and whose work it must be — were not public when this was written. Read them when the portal opens, before you accept feedback from anyone, us included, and follow them.

Within that boundary, what a rubric review does is narrow and specific: it reads the draft you paste against the published criteria for that award and hands back observations tied to your own sentences, so you can revise before your advisor's limited time gets spent on the first pass. It does not write for you, does not predict an outcome, does not verify facts outside the text, and does not judge eligibility. The scope is set out in how the scholarship essay review reads a draft.

For the fall fellowship documents with public rubrics today, start with the NDSEG application essays rubric and the GRFP personal statement rubric, both on the scholarship rubric shelf; the scholarship essay review page lists the rest. Two Quick Reviews are free each day, your first Full Review is $5 once per account, and Pro is $50 for ten Full Reviews.

More STEM fellowship guides

• GRFP, NDSEG, Hertz: apply to all three in one season
• NDSEG fellowship essays FY2027: two boxes, one deadline
• NSF GRFP 2027 cycle guide
• How to write a scholarship essay that fits the award

Sources

All Hertz Foundation, NDSEG, and NSF pages retrieved August 19, 2026. No sponsor fact on this page comes from a third-party site.

• Hertz Foundation — 2026 fellows announcement (1,493 applications, 119 interviewed, 54 second interviews, 19 fellows; record application volume)
• Hertz Foundation — 54 finalists named for the 2026 Hertz Fellowships (February 5, 2026; "more than 1,450 applicants")
• Hertz Foundation — Apply (2027 timeline; award "up to five years of funding, valued at up to $250,000")
• Hertz Foundation — Frequently Asked Questions (interview rounds; no required research project proposal; suggested essay lengths)
• Hertz Foundation — Who Can Apply (eligibility; the five published attributes)
• Hertz Foundation — Recommenders (form rather than letter; the assessed qualities)
• NDSEG — program home (cumulative totals since inception; FY2027 window)
• NSF — Graduate Research Fellowship Program and solicitation NSF 25-547
• NSF — GRFP awardee and honorable mention lists by award year

The percentages in this post are GradPilot calculations from the sponsor counts cited above, computed August 19, 2026. Sponsors publish counts; they do not publish these rates. Re-check every figure against the live sponsor pages before you rely on it — the portal governs.