White paper · 6 November 2026
Most accelerators do not add value. Pick the 60-day startup path that matches the company.
By Mike Sullivan. This is the public version of the 2026 acceleration paper. Product counts are the live 60-Day Founder school, not the draft totals in earlier notes.
For twenty years the default was a cohort accelerator: a few months, a small check, and 5% to 10% of the company. That trade made sense when a team and a rack of servers were the cost of starting. It is a bad default when one domain expert can ship software and the median program does not beat doing it alone.
What the research actually says
Baek and Hegde, in NBER Working Paper 35063, adapt a value-added method to roughly 750,000 US startups and 329 accelerators. Stronger companies sort into the famous programs, so alumni headlines mix selection with teaching. After that split, the median accelerator shows negative value-added against no program. The positive industry average is a handful of elite programs. The useful programs also kill weak ideas faster.
Chang and Assenova, in the Strategic Management Journal, study 6,723 startups in 281 accelerators. Accelerators amplify founders who already have startup experience. A year later, experienced founders in the same cohorts show much higher revenue and hiring growth than novices. Office hours and investor intros are not a curriculum. A first-time or domain-expert founder needs a sequence, not a networking hour.
Fitzgerald and Ray, using internal scores from a US venture accelerator (9,068 applications and 18,500 judge scores), find partner scores on the same file can differ by up to 25%. In solo screening, a shared affinity (school, gender, ethnicity) is associated with an 8% to 11% higher score, concentrated where there is no revenue yet. The bias fades when judges have to defend the score in a group. A pitch deck is a noisy instrument. A file of logged work is less so.
Giving up 5% to 10% before repeatable revenue also starts a clock. The company is expected to look like a Series A story in about two years. If it becomes a good, smaller business instead, the early dilution is still gone.
Four paths, and who each one is for
| Path | What it costs | Best fit |
|---|---|---|
| Elite cohort (YC, Techstars) | 5% to 10% equity | A repeat founder who can use the network and wants the signal |
| Virtual, equity-free program | A subscription, no equity | A founder building beside a job, outside a hub |
| Compute or corporate AI program | Varies, often credits rather than cash | A team that needs GPUs, not a sales coach |
| 60-day startup path (60-Day Founder) | $29 or $79 a month after a 7-day trial. No equity. | A domain expert or small team shipping software and keeping the cap table |
If the company needs a million dollars of hardware or a GPU cluster before the first dollar of revenue, apply to an elite or corporate program. If it does not, the equity is optional.
What the 60-day path includes
60-Day Founder is the self-directed path in the table. The live school, not the marketing draft, is:
- A 33-lesson core path inside a larger lesson library. The founder is not asked to finish every lesson to earn the credential.
- Nine workshops and twelve artifact types. Certification uses eight scored gate artifacts. A ninth workshop, the competitive map, is extra and does not raise the bar for people already certified.
- A pitch simulator (Dana Reyes) that checks what the founder says against the artifacts already saved.
- Small accountability groups, and a public credential at a verify link when the gates are passed.
- English, Spanish, and Portuguese.
The point of the file is to replace a subjective deck. Specialist micro-funds and screening models both read documents. A founder can map the artifacts into five folders: team and thesis, product, customer proof, capital efficiency, and the raise. Industry screens those investors already use, and which the book treats as targets rather than as this school's measured results, include a burn multiple under 1.5x, a Rule of 40 at or above 40, and CAC payback inside 12 months.
How to choose
- If you have already built a company and the prize is the elite network, pay the equity and apply there. The research says experienced founders are the ones who extract the mentorship.
- If you are a clinician, lawyer, operator, or engineer building vertical software, use the 60-day path. You need the sequence, and you do not need to sell 7% to get it.
- If you are building nights and weekends, a virtual equity-free program or the same 60-day path both avoid a move and a demo-day clock.
- If you are training models and the constraint is compute, take the credits from a corporate program. A curriculum will not replace GPUs.
The book is on Amazon
60-Day Founder: A Builder's Guide to Production AI is the English Kindle, $9.95. You can also read it inside the academy. The other two English Kindles are on the same author page.