The Portfolio-Count Test: Ask a Vertical Accelerator How Many Companies It Funded in Your Subsector

"Deep subsector expertise" costs nothing to write on a program page. A count of companies funded in your exact subsector, with a date and a source attached, costs something to publish: someone has to go count, and then the number can be checked. That cost is exactly why it means something.

Picture a climate founder building grid-scale storage. She finds a program with "climate" in its name, a sharp About page, and a fall deadline coming up. She reads the last two cohort lists and finds mostly consumer apps: carbon-tracking tools, a sustainable fashion marketplace, a composting subscription. All climate, technically. None of them have ever had to sit through a utility interconnection queue. The vertical label was accurate. It just wasn't about her.

That gap between vertical and subsector is what the portfolio-count test is for.

What to ask an accelerator before you apply

The test is one question with five parts. Ask the program for:

• The last two cycles. Recent cohorts, not the whole history. A program that backed three fintech infrastructure companies years ago and none since is telling you something.
• Your exact subsector. Not "fintech" but payments infrastructure, lending, or compliance tooling. Not "biotech" but diagnostics or therapeutics.
• The number. A count, not adjectives.
• The date it was counted. So you know whether it's current.
• Where it came from. A published cohort list, a portfolio page, a partner's own tally.

If they can answer all five, you've learned more than any About page will tell you. If they can't, you've also learned something (more on that below).

How to phrase it without sounding adversarial

First-time founders often worry this question sounds like an accusation. It doesn't, if you frame it around fit. On an intro call or in an application Q&A, try something like:

"We're in grid storage specifically. Over your last two cohorts, how many companies have you funded in storage or grid infrastructure? Is there a public list I can look at?"

That's it. You're asking because you want to know whether their mentors and investor network have seen your problems before. A good operator will appreciate the question, because founders who self-select for fit make better cohorts. This is education about what to ask, not advice on where to apply. The decision stays yours.

For a deeper walkthrough of reading cohort lists yourself, see how to check if a vertical accelerator has actually placed companies in your subsector.

Why a blank answer is still an answer

Here's the part most founders skip. When a program can't produce its own count, that's not a missing data point. That's the data point.

An accelerator knows who it funded. Its cohort lists exist. If it can't (or won't) tell you how many of them sit in your subsector, there are only a few explanations, and none of them are "we're secretly the best fit for you." Maybe the number is zero. Maybe nobody tracks it, which says something about how seriously subsector depth is taken. Either way, you can walk away on that signal without feeling like you're being unfair.

Accelerator Atlas runs on the same rule: a blank you can trust beats a number you can't. When we can't verify a figure against a program's own documents, we show a blank rather than fill it with a plausible guess. Every program in the directory carries one of three confidence tiers:

• Verified: checked against the program's primary documents.
• Secondary: sourced from a credible third party, not the program itself.
• Unverified: we couldn't confirm it, and we say so.

Each figure also shows its source and the date it was last checked. So when you're running the portfolio-count test, you can hold the program's answer up against what's on record and see whether they line up.

The vertical-versus-subsector problem shows up differently in every field. If you're in life sciences, the vertical umbrella trap covers how "biotech accelerator" can mean almost anything.

What Atlas's own numbers look like (and don't)

Since we're asking programs to show receipts, here are ours. Accelerator Atlas lists 105 US accelerator programs across 14 verticals. Of those, 102 are verified against the accelerator's own primary documents, not scraped summaries or hearsay. The other three are labeled for what they are. Programs that go quiet get flagged as dormant rather than deleted, because "this program hasn't run a cohort recently" is useful to know too.

One more thing worth saying plainly: Atlas takes no success-based or funding-tied fees. Revenue comes from flat subscriptions, newsletter sponsorships, and disclosed affiliates. We don't get paid more if you apply to one program over another, or if you get in, or if you raise. We have nothing riding on which program passes your test.

Once a program passes, compare the terms

The portfolio-count test is a filter, not a finish line. A program can have real depth in your subsector and still offer terms that don't work for you.

So after you've shortlisted the programs that answered with a dated, sourced number, look at the deal. How big is the check? How much equity does it take? Is the money on a SAFE (a simple agreement for future equity, which converts into shares at a later priced round) and does it include an MFN clause? If those terms are new to you, the Accelerator Atlas glossary and guides cover SAFE, MFN, and pro-rata without the jargon, so you don't have to decode them mid-application.

Side-by-side terms comparison lets you line up check size and equity across your shortlist in plain English. Two programs with strong grid-storage portfolios can ask for very different slices of your company, and you want to see that before you hand one over, not after.

Running the test during the fall wave

The fall wave is the year's second big deadline season, and shortlists get built fast. That's an argument for a quick filter, not for panic. The portfolio-count test takes one email or one question on a call. Programs that can answer it move forward. Programs that can't drop off. You end up applying to fewer places with more conviction, which is a better use of application hours than spraying a dozen portals.

Timing still matters, of course. Portal deadlines aren't always the real ones, and tracking accelerator deadlines by vertical walks through how to keep them straight without a spreadsheet meltdown.

Your next move

Browse the accelerator directory and filter by your vertical. For each program, check the confidence tier, the source, and the check-date. Run the portfolio-count test on the ones that look promising, then set deadline alerts for the programs that pass so you apply on time without rushing. Which of your shortlisted programs can actually tell you how many companies like yours it funded last cycle?