Y Combinator Acceptance Rate: What the Public Batch Data Actually Shows

Y Combinator’s public batch data supports a narrower conclusion than many acceptance-rate headlines imply. For two batches, the published application and funded counts can be divided directly. For two others, application totals stated as “over” a threshold allow only an upper bound. YC also publishes a detailed portrait of selected founders, but it does not publish individual rejection reasons. Keeping rates, cohort characteristics and admissions explanations separate gives applicants a more accurate picture.

What the published batch counts show

The most transparent public-data calculation divides the number of companies funded by the number of applications received:

• Summer 2022 (S22): YC reported 240 funded companies from 19,000 applications. The recomputed rate is 240 / 19,000 = 1.26%, or about 1.3%.
• Winter 2022 (W22): YC reported 414 funded companies from 17,000 applications. The recomputed rate is 414 / 17,000 = 2.44%, or about 2.4%.
• Winter 2023 (W23): YC reported 282 funded companies against “over 20,000 applications.” Using 20,000 as the stated threshold gives 1.41%, but the actual application total was higher. The defensible result is therefore under about 1.4%, not an exact acceptance rate.
• Winter 2024 (W24): YC reported 260 companies funded against “over 27,000 applications.” Because the application figure is not exact, dividing 260 by 27,000 would not produce an exact batch rate. YC itself describes W24’s acceptance rate as “under 1%.”

The S22 and W22 figures are recomputations from YC’s published application and funded counts. W23 and W24 require YC’s bounded wording to remain intact: under about 1.4% for W23 and under 1% for W24.

These are also batch-wide ratios, not the acceptance probability of an individual founder. They describe what happened in four historical cohorts; they do not establish the likelihood of acceptance from a future batch.

Why W23 and W24 can only be upper bounds

An exact acceptance rate requires an exact numerator and denominator. In W23, the funded count was 282, while the application count was only described as over 20,000. Every additional application beyond that threshold would make the resulting rate lower than the calculation based on 20,000.

That is why 282 / “over 20,000” means under 1.41%, which can be conservatively stated as under about 1.4%. It does not justify reporting a precise percentage such as 1.4%.

W24 has the same denominator problem. Its public inputs were 260 funded companies and over 27,000 applications. The application count cannot be treated as exactly 27,000, so the inputs do not support a more exact rate than the one YC published. Its official wording—“under 1%”—is the appropriate figure to retain.

The distinction matters because “over” is not a synonym for “approximately.” It identifies a threshold, not a final count. Presenting the threshold as exact would give the appearance of greater precision than YC’s public data supports.

What YC publicly says about who gets in

YC’s closest public substitute for a rejection profile is a description of the companies it funded. Its FAQ says that, on average, 40% of the companies funded in each batch are just an idea.

The Winter 2023 batch post provides a more detailed snapshot:

• 52% were accepted with only an idea.
• 77% had zero revenue before YC.
• 28% had raised money before YC.
• 2% had more than $50,000 in monthly revenue when accepted.

These are separate characteristics of the accepted cohort, not stages in a sequential admissions funnel, so they should not be added together. The figures also do not mean that every earlier-stage company had a particular admission advantage. They show the composition of one funded batch, not the probability of being selected.

YC’s FAQ separately reports that 7% of recent batches had more than $50,000 in monthly revenue when accepted. That broader recent-batch statement should not be substituted for W23’s batch-specific 2% figure or treated as an acceptance rate. In both cases, the subject is the revenue profile of selected companies rather than the percentage of applicants admitted.

Geography and sector provide more context for W23. The batch post says 86% of its founders lived in the Bay Area. Its largest published sector shares were 54% in B2B/Enterprise SaaS, 17% in DevTools and 12% in Fintech. Those are descriptive features of that cohort, not published location or sector requirements. They also do not create a Hong Kong or wider Asia-Pacific acceptance rate; the figures cited here apply to the batch as a whole.

For an early-stage founder, the central message is that YC routinely funds companies with only an idea and companies with no revenue. At the same time, the presence of previously funded founders and a small revenue-producing group shows that the funded cohort was not uniformly pre-revenue. The data describes a range of circumstances rather than a single mandatory profile.

Why there is no public rejection profile

YC does not publish rejection reasons, a rejection taxonomy or an internal scoring rubric explaining why one application was selected over another. Its public batch pages describe selected cohorts and application volumes, but they do not convert those descriptions into a list of reasons for non-selection.

The FAQ is equally direct about feedback: YC says it does not provide feedback on application results unless a founder is invited to interview. It attributes that policy to the volume of applications its team must process.

That leaves an important limitation. Aggregate statistics can show that many funded companies were at the idea stage, had no revenue or lived in a particular region. They cannot reveal why a particular rejected application was weaker, whether one missing attribute mattered, or how applicants were ranked. An accepted-cohort profile is not a model of the unsuccessful application pool.

The absence of published rejection reasons also means applicants should be cautious with confident explanations for a non-selection. Statements that point to a specific factor as the cause go beyond what YC has disclosed.

What founders can do with the public data

The actionable alternative begins with accepting the limits of the evidence. Founders can use the batch statistics to understand the kinds of companies YC has funded, but they should not treat them as a scoring checklist. For example, W23’s 77% zero-revenue figure does not establish that zero revenue is required for acceptance, just as its 86% Bay Area figure does not establish a location preference that every applicant must match.

Reapplication is a more concrete route. YC’s FAQ explicitly says applicants can apply more than once and reports that, in a typical batch, about half of the companies applied multiple times before being accepted. The W23 batch was slightly above that typical share: 59% had applied more than once.

That statistic should likewise be read carefully. It describes the prior application histories of accepted companies; it does not prove that reapplying increases any individual founder’s probability of acceptance. It simply confirms that multiple applications are possible and common among the companies eventually funded.

YC’s feedback policy gives applicants a second clear boundary. Before an interview, the published policy provides no feedback on the application result. If invited, the interview is therefore the point at which a founder can seek that feedback. Reapplying and preparing to use any interview feedback are the concrete alternatives supported by YC’s own public material.

The defensible summary is therefore straightforward: the public figures produce about 1.3% for S22, about 2.4% for W22, under about 1.4% for W23 and under 1% for W24. They also show that YC often funds very early-stage founders, that repeat applications are common among accepted companies, and that no public rejection reasons are available. The data can inform an application strategy, but it cannot supply a guaranteed outcome or explain every unsuccessful application.