01
Discovery used to cost millions.
Learning whether a company would work required a year of engineering, infrastructure, and payroll before any real signal arrived. That expense forced the industry to invest against judgment rather than evidence.
Cupel Studio · A venture studio for the evidence era
Ninety days.
One reading.
We build B2B software companies to a ninety-day commercial reading, then let the data decide what happens next.
Ninety-day cadence
15–25% common, set at intake
Founder holds the seat
Delaware C-corp
01 · The thesis
Venture is a prediction business in a world that no longer requires prediction.
02
Venture is a guessing craft.
The deck, the warm intro, the pattern match, the founder-first thesis, the portfolio math built around ninety percent failure. Instruments for allocating capital in the absence of proof.
03
Proof is now affordable.
A production product ships in weeks. Infrastructure is credited. Ninety days and roughly five thousand dollars produce a genuine commercial reading. When measurement is cheaper than judgment, the case for judgment weakens.
“We are not better guessers. We have stopped guessing.”
Includes what would prove us wrongThe full thesis →
02 · The cadence
Ninety days,
one reading.
Long enough for an enterprise buyer to answer. Short enough for the answer to stay affordable.
The clock · Day 0 → Day 90+Click any stage
Tap a stage for the definition:
Days 1 – 3001
Build
A production product in front of one named enterprise buyer, with contract value on the table. Shipped, not demoed.
Cash cost: near zero
Days 31 – 6002
Sell
Studio-run outbound against a defined account list. The founder holds every buyer conversation and every close.
Named accounts, priced offers
Days 61 – 9003
Read
Revenue, slope, retention, and objection quality against a single threshold. $100K ARR is the point at which the question stops being subjective.
One threshold, one reading
Day 90 · Report honestly
Four readings. The data selects.
03 · Day 90
All four readings, published.
The industry records binary outcomes because a prediction can only be right or wrong. We publish the full distribution.
GO— seed-ready
Revenue and slope sufficient to underwrite the next round. Warm introductions to lead investors. Founder runs the process.
GROW— commercial, not venture-shaped
A working business without seed-round trajectory. Continue and re-test in a quarter, or operate for cash flow under joint ownership.
PARK— signal early
The proposition holds; the market or the metric is not ready. Low-cost hold with a set re-evaluation date.
KILL— negative reading
The wedge did not land. You retain everything: the company, the code, the customers, the domain. No clawback.
“A kill is a completed experiment. The expensive mistake is keeping the company alive to protect the original bet.”
04 · The infrastructure
Cheap enough to run many.
A method built on evidence only holds if the evidence stays affordable to collect.
Non-dilutive infrastructure · per company · Day 1
up to$500Kin stackable maximums
The higher tiers normally require venture backing or a referral. As a registered provider, we place them into every company on Day 1. Typical amount actually deployed across the 90 days: $50–75K per company, weighted to compute and outbound. The rest is available as the company grows into it.
Provider 01
Microsoft Founders Hub
up to $150K
Azure + Azure OpenAI + GitHub Enterprise
Provider 02
AWS Activate
up to $100K
Portfolio tier, via our Provider ID + Bedrock
Provider 03
Google for Startups
up to $350K
AI-First track + Vertex + Gemini
Provider 04
NVIDIA Inception
$100K+
for compute-heavy builds
Stacked provider maximums · verified quarterly · actual allocation varies by tier
05 · The operators
Built by operators, held by an investor.
The method requires the discipline of shipping and closing paying customers. It cannot be run from a term sheet.
Founder & Managing Partner
Jack Gierlich
i
Fifteen years of operating.
One company built as a founder and sold; two others sold as a senior operator in VP and CMO roles. Regulated healthcare, where a single enterprise sale takes a year and every buyer conversation has to be earned.
ii
Both sides of the table.
Several years sourcing early-stage healthcare software into venture pipelines. Dozens of companies evaluated on decks and narrative because the evidence did not yet exist to evaluate them on.
Three prior exits
2020
Acquired
Regulated-healthcare go-to-market. Senior operator through the transaction.
2021
Acquired · public buyer
Clinical trial operations software. VP-level operator through the public acquisition.
2025
Acquired
Patient-matching commercial, acquired while CMO. Interim CEO post-acquisition.
Additional operating partners announced on joining.
06 · Intake
Three ways to enter Cohort 01.
Door A
A
Adopt
You have a product with early signal.
We attach the studio’s sales engine, capital network, and operating cadence to what you have already built. Your idea, your code, your customers.
Door B
B
Assemble
You are a founder without a hypothesis.
Select a pre-scoped wedge: named buyer, sized ACV, a stated reason it should work. Validated by the studio before you write a line of code.
Door C
C
Pair
You are half the founding team.
A technical founder without a commercial counterpart, or the reverse. We introduce a match inside the cohort with equity documented before Day 1.
07 · The economics
Every term, before the term sheet.
The studio is compensated on a positive commercial outcome. A negative reading returns ninety days.
§ Terms7 clauses
- Studio equity
- 15–25% common, minority. Adopt at the lower end, Assemble at the higher. Documented before Day 1.
- Build-value SAFE
- Studio build, GTM, and IP contribution converts to preferred at your priced seed, on your investors’ terms. Absent a round, it does not convert.
- Build-cost recoup
- Paid only from the proceeds of a priced round, never from operating cash.
- Capped revenue share
- Small, capped, activated only after a priced round or sustained cash flow.
- Founder cash cost
- None. Build stack, credits, sales team, and operators are compensated in studio equity.
- Cap table
- Institutional from Day 1: founder common, studio minority preferred, single option pool.
- Founder control
- Board majority and CEO seat retained. On termination, the company and domain remain with the founder.
§ 01
§ 02
§ 03
§ 04
§ 05
§ 06
§ 07
Worked SAFE examples at different raise sizes: /safe-terms.
08 · Distinctions
What the studio is not.
- Not an accelerator
No curriculum, no batch, no demo day. A cadence with a reporting date.
- Not an agency
The studio does not build on contract. Every build results in a co-owned company.
- Not a fund
The studio operates inside the company. A fund follows the reading, on institutional terms.
- Not a coaching program
Companies that require coaching more than commercial validation are not a fit.
- Not permanent
Every engagement resolves at Day 90. No portfolio limbo.
09 · Fit
Who the studio is for.
Apply if
- You would trade a two-year maybe for a ninety-day answer.
- You can build, or you will pair with a technical counterpart.
- You are pursuing venture-scale B2B software with a defined buyer and a real ACV.
- You can receive a negative reading without treating it as a verdict on yourself.
Don’t apply if
- You want conviction in your vision more than a test of it.
- You want capital without an operating partner attached.
- You are building a lifestyle business, an agency, or a consultancy.
- You are pre-idea, non-technical, and unwilling to take a scoped wedge.
10 · The record
Cohort 01, in flight.
Every company appears here at Day 90 with its reading, kills included.
Slot 01
Awaiting Day 90
Company, buyer, reading, and outcome published on report.
Slot 02
Awaiting Day 90
Company, buyer, reading, and outcome published on report.
Slot 03
Awaiting Day 90
Company, buyer, reading, and outcome published on report.
11 · Questions
On the record.
15–25% common, minority. Adopt at the lower end, Assemble at the higher. Documented before Day 1.
None. Build stack, credits, sales team, and operators are compensated in studio equity. The founder contributes product and time.
The reading determines the path. GROW if revenue is real but not venture-shaped. PARK if the signal is early. KILL if the wedge did not land, in which case the founder retains the company in full.
The founder. Board majority and CEO seat retained, domain in the company's name. The studio holds a minority preferred stake via the build-value SAFE and an observer seat.
The studio prepares the evidence package and opens warm introductions to lead investors. The founder runs the process and closes.
Cohort 01 · Reviewed weekly · Reply within 7 days
Ninety days
to an
answer.
A short application. An operator replies within seven days.
Apply →