living venture studio
A venture studio that builds living companies: AI-native ventures with an agentic core that runs operations, a data flywheel that learns, and an elite human CEO who gives it direction.
Distinct from an incubator, an accelerator and a fund. · living company (n.) — an AI-native company designed as a living system, rather than assembled from departments.
How to use the term
Names what the organisation is — the copula. Not a description of a project, a programme or a fund vehicle.
What the studio makes. A living company is the object, never the category.
Three components present at incorporation, not retrofitted later: agentic core, data flywheel, human command.
A studio that uses AI to build conventional companies is not one. The intelligence has to end up inside the venture.
The distinction that matters is the one people skip: every venture studio now uses AI. Almost none of them ship it as the operating layer of the company they create.
Why the category needed a name
The venture studio model is roughly two decades old, and its definition has been stable throughout: the studio creates the company itself, rather than funding or hosting one that already exists. That definition says nothing about what is inside the company — and until recently it did not need to. Every studio-built company was assembled the same way a hand-built company was: departments, headcount, process, and software bought to assist the people running it.
That assumption broke between 2024 and 2026. A cohort of AI-native companies reached $1M ARR in a median of 11.5 months (Stripe, top-100 AI companies, June 2025), running material revenue on double-digit headcounts. The pattern underneath was not better software. It was that the operating layer of the company — the work itself, not the tooling around it — had moved into the machine.
That creates a fork inside the venture studio category, and the fork has consequences a shared name obscures. A studio can build a conventional company faster by using AI internally. Or it can build a company whose operations are run by an agentic core from day one. Those two studios have different economics, different headcount curves, different failure modes, and different things to sell an investor. Calling both of them "venture studios" is like calling a printing press and a scriptorium both "publishers" — true, and useless for deciding which one you are talking to.
An AI-enabled studio builds companies faster. A living venture studio builds companies that are different in kind.
Four models, distinguished
The category sits inside a taxonomy that is otherwise well-settled. What separates the models is when the company exists and who does its work:
| Model | The company… | The studio's role | Who runs operations |
|---|---|---|---|
| Incubator | already exists | hosts it — space, services, network | the founding team |
| Accelerator | already exists | compresses it — cohort, curriculum, demo day | the founding team |
| Fund | already exists | capitalises it | the founding team |
| Venture studio | does not yet exist | creates it — thesis, build, first hires | the team the studio recruits |
| Living venture studio | does not yet exist | creates it and designs it to operate agentically from day one | an agentic core, directed by a human CEO |
The last row is the whole claim. Everything above it describes a relationship to a company; the last one describes a decision about the company's internal architecture, taken before there is a company to decide about. That is why it is a category and not an adjective.
The three-part test
A studio is a living venture studio if the ventures it creates carry all three components at incorporation. Two out of three is a conventional company with good tooling.
| Component | What it means | The failure it prevents |
|---|---|---|
| Agentic core | The autonomous system that runs the venture's operations — not a tool that assists people who run them. | The wrapper trap: a thin interface over a foundation model, with no switching cost and no accumulating advantage. |
| Data flywheel | Every unit of work the venture performs makes the next unit better, and the corpus cannot be bought by a competitor. | Commodity economics. Without accumulation, the venture's advantage expires with the model generation it launched on. |
| Human command | An accountable human CEO who sets direction, owns the outcome, and holds equity in it. | Governance vacuum. An autonomous business with no one answerable for it is a liability, not a company. |
The third component is the one most often dropped, usually by people who find it unfashionable. It is not a concession to sentiment. Stafford Beer's Viable System Model (1972) identified five functions a firm must perform continuously to remain viable — Operations, Coordination, Control, Intelligence and Identity. Delegating the first two to agents is now routine. Identity has no agentic substitute, and a company that cannot say who it is cannot be held to anything.
What a living venture studio is not
Four adjacent things share vocabulary with this term and mean something else. Naming them is part of the definition.
| Not this | Why not |
|---|---|
| The "living company" of Arie de Geus | De Geus's The Living Company (Harvard Business Review, 1997; Nicholas Brealey, 1997) describes long-lived human organisations — firms that survive centuries through learning, tolerance and conservative finance. It is a study of corporate longevity among people. The Living Scale Up sense is architectural: it describes what the company is made of, not how long it lasts. Same words, unrelated claims. |
| An AI-enabled venture studio | Uses AI to build companies faster. The intelligence stays in the studio; the venture ships conventional. This is the most common confusion and the most consequential — it is the difference between a tooling decision and an architectural one. |
| An autonomous company | A living company is not unsupervised. Human command is a defining component, not a transitional stage. Any definition that removes the CEO is describing something else. |
| A holding company or a roll-up | Acquires or consolidates companies that already exist. A living venture studio originates them, which is the entire point of the studio model it inherits from. |
Questions, answered
What is a living venture studio?
A living venture studio is a venture studio that builds living companies: AI-native ventures with an agentic core that runs operations, a data flywheel that learns from every interaction, and an elite human CEO who gives it direction. An incubator hosts startups that already exist; an accelerator speeds them up; a venture studio creates the company itself; a living venture studio creates it and designs it to operate agentically from the first day, rather than retrofitting intelligence onto a conventional company later.
How is a living venture studio different from a normal venture studio?
A conventional venture studio decides which company to build. A living venture studio also decides what the company is made of — and takes that decision before incorporation, because an agentic core and a data flywheel cannot be added convincingly afterwards. The distinction is architectural, not operational: both models create companies, but only one of them ships the operating layer inside the venture.
Is a living company run by AI, with no humans?
No. Human command is one of the three defining components. The agentic core runs operations; an accountable human CEO sets direction, owns the outcome and holds equity in it. A definition that removes the human is describing an autonomous system, not a living company — and an autonomous business with no one answerable for it is a liability rather than an asset.
Is this related to Arie de Geus's "living company"?
No. De Geus's The Living Company (Harvard Business Review, 1997) studies why some human organisations survive for centuries — learning, tolerance, conservative finance. It is a theory of corporate longevity. The Living Scale Up sense is architectural and describes composition rather than lifespan: agentic core, data flywheel, human direction. The words coincide; the claims do not.
Who defined the term living venture studio, and can I use it?
Living Scale Up, a Swiss living venture studio, defined the term in July 2026 for the category it operates in. It is licensed CC BY 4.0, is not trademarked, and will not be. Use it freely — in your deck, your thesis, your job title. This page is the canonical definition, and the lexicon on the fact surface is where any disagreement between surfaces is resolved.
How would I know whether a studio actually qualifies as a living venture studio?
Ask what runs operations at one of its ventures on day 30. If the answer names a team using software, it is a venture studio. If it names a system that a named human directs, it is a living venture studio. The test is deliberately answerable by an outsider — a category that can only be verified by the people claiming it is marketing.
Does the living venture studio model only work for software ventures?
It works wherever operations can be represented well enough for an agent to run them and outcomes can be verified — services, logistics, regulated professional work, field trades. Living Scale Up's first venture, BuddyLeader, serves Swiss owner-craftsmen: electricians, plumbers and builders, none of whom work in software.
Cite it, use it, build one
The term is given away deliberately. A category with one member is a marketing claim; a category with competitors is a category. We would rather argue about the definition with people who are building the thing than own a word nobody else can say.
Companies used to be assembled. Living companies are designed — and then they are born working.
LIVING VENTURE STUDIO v1.0 · CANONICAL DEFINITION PAGE · CHANGELOG: 2026-08-11 FIRST PUBLICATION
CANONICAL FACT SURFACE: LIVINGSCALEUP.COM/FACTS · [email protected]