YC's 2015 to 2019 batches made 43 unicorns. Techstars' made 7.
Published 21 October 2026
Y Combinator and Techstars took in a similar number of companies between 2015 and 2019: 1,324 in YC's ten batches, 1,484 across 144 Techstars program sessions in 58 programs. Seven to eleven years later, 43 of the YC companies are unicorns. Techstars counts 7 of its own.
That makes a YC company from those years about 7 times as likely to have become a unicorn (95% interval 3 to 15). The same gap shows on every other outcome we checked except exits.
| 2015 to 2019 | YC | Techstars |
|---|---|---|
| Companies | 1,324 | 1,484 |
| Unicorns | 43 (3.2%) | 7 (0.5%) |
| Raised $100M+ | 113 (8.5%) | 21 (1.4%) |
| Series A or later | 340 (25.7%) | 153 (10.3%) |
| Total raised | $87.4B | $11.7B |
| In Crunchbase | 78% | 50% |
| Exited | 22% | 20% |
"In Crunchbase" means the company has a funding round in our Crunchbase copy. "Exited" is acquired or public by YC's status field, and Techstars' own exit flag.
Every year looks the same
Share of each cohort that reached Series A or later:
| Cohort | YC | Techstars |
|---|---|---|
| 2015 | 22% (47/214) | 12% (23/194) |
| 2016 | 29% (64/224) | 8% (18/217) |
| 2017 | 23% (56/241) | 12% (35/291) |
| 2018 | 30% (82/277) | 12% (46/378) |
| 2019 | 25% (91/368) | 8% (31/404) |
Techstars' seven
Techstars flags seven companies from these years as $1B companies: GlossGenius (New York, 2015), Preply (Berlin, 2015), Self (Austin, 2015), Alloy and Chainalysis (both from the Barclays program in New York, 2015), Twelve (the Equinor energy program, 2018) and Veho (Boulder, 2019). Five of the seven come from the 2015 cohort. Crunchbase tags only one of them, Chainalysis, as a unicorn. We used Techstars' count because it is the larger one.
The 43 YC companies are the ones Crunchbase tags, among them Brex, Ramp, Rippling, Faire, Retool, Vanta, Scale, Replit, Razorpay, Meesho and Groww.
What this does and does not show
Exits are close: 22% of the YC companies are acquired or public by YC's status field, and Techstars marks 20% of its companies as exited. The difference is in the companies that went on to raise $100M or more.
These are outcomes, not a measure of what either program adds. YC picks from a very large applicant pool and invests in every batch company itself. Techstars ran about 29 program sessions a year in this period, some with a corporate partner such as Barclays or Equinor, and the mix of programs changes the mix of companies. The data cannot separate selection from treatment.
Method
- YC list: every company in YC's public company directory with a batch from Winter 2015 to Summer 2019, including inactive and acquired ones. 1,324 companies.
- Techstars list: every company in the Techstars portfolio search marked as an accelerator company, with a first program session from 2015 to 2019. 1,484 companies. We do not know whether the Techstars directory lists companies that shut down. If it leaves them out, Techstars' shares above are too high, not too low.
- Matching: both lists were joined to the FundingWatcher startup file (Crunchbase) by registered domain. Stage, total raised and the unicorn tag come from Crunchbase.
- Shares: stage shares use the whole cohort as the denominator, so a company with no Crunchbase match counts as not having reached Series A.
- Unicorns: "unicorn" for YC is Crunchbase's tag; we did not check valuations, and a YC company whose website differs from its Crunchbase domain is missed, so 43 is a floor. For Techstars it is Techstars' own flag.
- Interval: for the ratio of unicorn rates, on the log scale.
Directories read on October 21, 2026; Crunchbase data as of October 19, 2026.
Data
companies.csv: all 2,808 companies with accelerator, cohort year, website, Crunchbase stage, total raised and both unicorn flags. Accelerator Series A and closure rates for a wider set of programs are on the accelerator success rate page.