"Run an accelerator" is how this work gets described, and it is wrong often enough to be worth correcting with numbers.
We have delivered 27 programs across seven countries, each commissioned by a named institution. Sorting them by what the institution actually bought produces a different picture from the one the word "accelerator" suggests.
Only a quarter of the programs are accelerators
| What was commissioned | Programs |
|---|---|
| Acceleration | 7 |
| Global market entry | 5 |
| US market entry | 3 |
| Open innovation | 3 |
| Inbound (bringing foreign founders in) | 2 |
| Conference and IR preparation | 3 |
| Other (YC prep, mentoring, ecosystem networking, market entry) | 4 |
Seven of 27 are acceleration. Eleven are some form of market entry — taking a country's founders to a market where they have no network, or bringing foreign founders into the commissioner's own ecosystem. That is a different product with a different success measure: not how many startups graduated, but how many landed a customer, a partner or an office somewhere new.
Who signs for it
Sorted by the kind of institution that commissioned the work:
| Commissioning institution | Programs |
|---|---|
| Government agency or city | 16 |
| Bank | 5 |
| University | 2 |
| Corporate | 1 |
| Other | 3 |
The public sector commissions most of it — ministries, national agencies, city governments and prefectures. Banks are the second largest buyer, which surprises people who expect corporates there. A bank commissioning a startup program is usually solving two problems at once: it needs the innovation, and it needs to be seen supporting the national ecosystem it operates in.
Cohort size is not the measure, but it is the question
Cohort sizes are recorded for 19 of the 27 programs. They run from single-digit cohorts to 549 founders in one program. The large ones are open calls with a funnel behind them; the small ones are curated tracks where each company gets scheduled introductions.
Both work. What does not work is running the second with the budget and staffing of the first, which is the most common way a commissioned program disappoints everyone involved.
What an institution should ask before commissioning
- What is the outcome you can report? Graduation certificates are not one. Signed pilots, follow-on funding and companies that opened an office are.
- Who is the buyer for the startups you select? If the program does not have corporates or agencies attached who will run a pilot, the cohort ends where it started.
- Is the operator in the market? A program designed elsewhere and flown in spends its first month learning what a local team already knows.
- What happens in month four? Most programs end at demo day. The companies that need follow-on support get it in the quarter after, or not at all.
Method: 27 programs from 2080's internal delivery record, each with a named commissioning institution and market. Rows we could not verify internally are excluded from the record and from these counts. Cohort sizes are those recorded at the time of delivery, available for 19 of 27.