A venture studio builds companies instead of investing in them. That single sentence is the whole model, and almost everything else — the capital structure, the equity splits, the hiring, the failure modes — falls out of it as a consequence.
This page is the guide we wish had existed when we started. It covers what a studio actually is, how studios make money, the four structural decisions you have to make before you incorporate anything, what the studio owns centrally versus what each venture owns, how AI has changed the arithmetic, and the specific ways studios die. It is long because the short versions are all wrong.
What a venture studio actually is
A venture studio — also called a startup studio, a company builder, or a venture builder — is an organization that originates companies internally, staffs them, funds them, runs them through a shared set of operating systems, and then hands each one to a dedicated CEO who owns it from there.
The defining feature is not the money. It is the work. An investor writes a check and waits, holding a portfolio of bets whose outcomes they influence at the margins. A studio does the building: the idea, the first version of the product, the first customers, the brand, the hiring, the pricing. The studio is a participant, not a spectator.
That difference cascades into everything else. Because the studio does the work, it takes a much larger ownership stake than an investor would. Because it takes a larger stake, it can only run a handful of ventures at once rather than dozens. Because it runs only a handful, each one has to be chosen carefully rather than sprayed across a portfolio. And because each one is chosen carefully, the studio's judgment about which problems are worth solving becomes the single highest-leverage asset in the business.
There is a second family of studios worth naming: those that acquire rather than originate. Buying a small software product with existing payers and installing a new operator is structurally the same model with the discovery risk pre-paid. Most serious studios do both, because the two paths de-risk each other — an acquisition can carry a studio through the long middle of an originated venture.
Studio vs. accelerator vs. incubator vs. VC
These four get used interchangeably and they are not remotely the same thing. The differences that matter are: who has the idea, who does the work, how much equity changes hands, and how many companies run at once.
| Model | Idea comes from | Who builds | Typical equity | Concurrency |
|---|---|---|---|---|
| Venture studio | The studio | The studio, then an installed operator | Large — often a third to a majority | Two to six |
| Accelerator | The founders | The founders | Small — typically single digits | Dozens per cohort |
| Incubator | The founders | The founders, with space and services | Small or none | Many, open-ended |
| Venture capital | The founders | The founders | Minority, priced by round | Twenty to a hundred |
The useful way to read that table is as a trade between ownership and volume. Accelerators and funds solve the same problem — how do I get exposure to a lot of outcomes — by owning a little of many companies. A studio makes the opposite bet: own a lot of very few, and earn that ownership by doing work nobody else will do.
The bet only pays if the studio's involvement genuinely raises the odds. A studio that takes forty percent and adds nothing an accelerator wouldn't have added is simply an expensive accelerator, and the market corrects that quickly. Everything in the rest of this guide is, in one way or another, about earning the stake.
How venture studios make money
There are four revenue paths, and healthy studios usually run two or three of them at once.
1. Equity in the ventures
The headline mechanism and the slowest. The studio holds a large stake in each company and realizes it on an exit, a recapitalization, or a secondary sale. This is where the outsized returns live, and it is also why studios need a way to survive the five to eight years before any of it arrives.
2. Distributions from profitable ventures
Underrated, and the reason we prefer businesses with a path to profit that doesn't depend on raising again. A venture that throws off cash pays the studio a dividend on its ownership long before any exit event. This turns the studio from a fund that must eventually liquidate into a holding company that can compound. If you only take one structural idea from this page, take this one.
3. Services or a management fee
Some studios charge ventures for shared services — design, engineering, recruiting, finance — either at cost or at a markup. Some raise a fund and take a standard management fee on it. Both smooth cash flow. Both also create a subtle misalignment: the studio starts optimizing for billable work rather than for venture outcomes. Use it as a bridge, not as the business.
4. Selling ventures early
Building a company specifically to sell it within two or three years to a strategic buyer. This is a real and legitimate strategy, particularly for studios embedded in a specific industry with obvious acquirers. It requires a different kind of venture selection than the compounding path, and mixing the two inside one portfolio muddles both.
The practical question for a new studio is not "which of these is best" but "which one pays the bills in year two". Equity pays in year seven. Something has to cover the gap.
Choosing a capital structure
Three structures dominate, and picking the wrong one is close to unrecoverable — it determines who you answer to for the next decade.
Balance sheet / self-funded
The studio funds ventures from its own cash, generated by an existing business, prior exits, or the ventures themselves. Maximum control, maximum patience, no external clock. The constraint is obvious: you can only build as fast as you can fund. Most studios that survive their first five years start here, because nothing else gives you the freedom to be wrong twice.
Studio fund (two-entity)
An operating company (the studio, its people, its playbooks) alongside a fund vehicle that capitalizes the ventures. LPs get familiar fund economics; the studio entity holds the founder-equity stakes. This is the standard for scaled studios and it is genuinely the right answer once you have a track record. It also imports a fund's timeline — a ten-year life with pressure to mark up and exit — into a model whose best outcomes often want to be held.
Holding company
One entity owns all the ventures permanently. No fund life, no forced exits, dividends flow up. It fits the "durable businesses that pay for themselves" thesis better than any other structure. Raising against it is harder because you are asking investors for permanent capital, which is a much narrower market. Berkshire is the archetype and also the reason people underestimate how long it takes.
A caution worth internalizing: the structure has to match the ventures. A holdco full of businesses that need eight-figure growth rounds will starve. A ten-year fund full of profitable, slow-compounding SaaS will be forced to sell good companies at bad moments. Decide what kind of company you intend to build first, then pick the wrapper.
Equity: the hardest problem in the model
Every studio eventually has the same argument, usually around venture number three. The studio contributed the idea, the capital, the first version of the product, and the first ten customers. The operator is going to spend the next six years of their life running it. How do you split that?
The failure mode on one side is a studio that keeps so much that no operator worth hiring will take the job, and the ones who do take it behave like employees. The failure mode on the other side is a studio that gives away so much it can't fund the next venture, which is the whole point of being a studio rather than a company.
What we've found is that a single fixed formula is the wrong shape for this problem. The cases are genuinely different:
- A founder bringing a working product with paying customers has already absorbed the risk the studio would otherwise take. They should hold correspondingly more.
- An operator stepping into a venture the studio originated and validated is taking a much smaller leap and joining something that already works.
- An operator joining pre-revenue to build from a thesis is closer to a co-founder and should be treated as one.
Forcing all three into one number produces a bad deal for somebody, and the person it's bad for figures it out eventually. What should stay constant is the principle: whoever runs the company holds meaningful ownership in it, and the vesting is tied to durable growth rather than to a raise or a headcount target.
Two mechanics that make the ranges workable regardless of where you land: vest over four years with a real cliff, so a mismatch in year one doesn't permanently encumber the cap table; and put a buyback at fair value on departure, so a venture that outlives its first CEO isn't carrying a large passive holder forever. Neither is exotic. Both get skipped constantly, and both are painful to add retroactively.
How we structure operator roles goes into what the job actually involves on our side.
The studio operating system
The thing that makes a studio more than a collection of companies is the set of assets every new venture inherits on day one. If venture four starts from the same blank page as venture one, you don't have a studio — you have a portfolio with extra overhead.
Five systems are worth centralizing, roughly in the order you'll need them:
- Discovery. A written process for going from a suspected problem to a validated one: who to interview, what to ask, what evidence counts as a yes, and — critically — what evidence counts as a no. Studios drown in ventures that were never killed early enough.
- Brand and onboarding. A repeatable path from nothing to a credible product identity and a first-run experience that activates users without a human in the loop. This is the single most compressible cost across ventures.
- Engineering and data. Shared architecture, deployment, and analytics so that unit economics are visible from week one rather than reconstructed in year two. The studio should be able to see every venture's retention curve in the same format.
- Growth. The acquisition motions that have worked before — partnerships, content, outbound — documented well enough that a new venture can run them without rediscovering them.
- Operator hiring. A real vetting process, run repeatedly, with a bench you're building before you need it. Studios consistently underestimate this one and then spend nine months searching under pressure.
The discipline is to write these down after each venture, while it's fresh, and to treat them as living documents rather than artifacts. Our five playbooks are the version of this we run.
How AI changed the math
The venture studio model is roughly thirty years old. It was, for most of that time, expensive — the reason so few existed is that building the first version of a company cost more than most people could fund repeatedly.
That constraint has substantially weakened, and it changes three things:
The cost of a first version collapsed. What was a two-quarter build with a team is now, for a large class of software, a matter of weeks. The studio's scarce resource shifted from engineering capacity to judgment about what to point it at. This is the core argument for an AI startup studio as a distinct category: the model works far better when the build step is cheap.
The right number of concurrent ventures went up — but not as much as people think. Building is cheaper; distribution is not. Customer acquisition, trust, support, and compliance did not get cheaper, and those are where ventures actually stall. Studios that took the cost collapse as a licence to run twelve ventures at once mostly discovered they had twelve distribution problems.
Durability got harder to assess. If a product takes weeks to build, it takes a competitor weeks to copy. The defensibility moved to the things that were always slow: proprietary data, regulatory position, workflow entrenchment, brand, and genuine distribution. A studio operating today has to underwrite for a moat that isn't the software, because the software isn't one.
That last point is why our own filter is weighted toward durable problems and recurring revenue rather than technical novelty. We've written more on this in the journal.
Why venture studios fail
The five patterns, in roughly the order they kill people:
Running out of runway before equity matures. The most common death by a wide margin. Equity is a seven-year instrument and rent is monthly. Studios that don't solve the year-two cash problem — through profitable ventures, services, or a large enough balance sheet — don't get to find out whether their thesis was right.
Too many ventures at once. Studio attention is the scarce input and it does not divide cleanly. Three ventures with real attention beat eight with a check-in call. The temptation to add one more is strongest exactly when the current ones are hardest, which is precisely when you shouldn't.
Never actually handing over. A studio that stays in the driver's seat has not built companies, it has built departments. The handoff has to be genuine: the operator owns the P&L, makes the calls, and lives with them. Studios that can't let go cap every venture at the founder's personal bandwidth.
Hiring operators too late. The search takes months and the pressure to fill it fast rises with every week the venture drifts. Build the bench before the seat is empty.
Killing nothing. A studio's discovery process is only as good as its willingness to act on a no. Ventures that should have been stopped at month four and instead limped to month twenty consume the attention that venture five needed. Write the kill criteria down before you start, because you will not be objective later.
Your first twelve months
If you are starting one, this is the sequence that avoids the most expensive mistakes.
Months 1–2: decide what you are. Write down the thesis — the specific kinds of problems you'll work on and, more usefully, the ones you won't. Pick the capital structure that matches. Answer the year-two cash question explicitly and in writing, before it becomes urgent.
Months 2–5: run one venture end to end. Not three. One. The purpose of the first venture is partly the venture and mostly the process: you are trying to find out what your discovery step is actually missing, and you cannot learn that in parallel.
Months 4–8: write the playbooks as you go. Every time you solve something — a pricing page that converted, an onboarding sequence that worked, an interview script that surfaced a real no — write it down that week. Retrospective documentation is always thinner and always later than you plan.
Months 6–10: start the operator search before you need it. Meet people with no role to offer. The best operators are not looking, and the relationship takes longer than the hiring process.
Months 9–12: hand over, then start venture two. The handoff is the real test of whether you built a studio or a company. If venture two can start with more than venture one had — a playbook, a template, a pipeline, a known-good architecture — the model is working.
The unglamorous truth is that a venture studio compounds only if each venture makes the next one cheaper. Everything above is in service of that one sentence.