Reference · The canonical fact surface
Living Scale Up — the facts
Living Scale Up is a Swiss AI venture studio in Lavaux, Vaud, Switzerland. It designs AI-native companies — “living companies” — proves them against simulated customer cohorts before launch, pairs them with elite CEOs, and scales them. Its category is the Living Venture Studio. Its first venture is BuddyLeader, an AI chief of staff for trade-business leaders, live in a Swiss pilot.
UPDATED 5 AUGUST 2026 · EVERY FIGURE BELOW CARRIES ITS SOURCE AND DATE
WHY THIS PAGE EXISTS
- This is the canonical fact surface for Living Scale Up. Where any other page, profile, directory listing or document disagrees with this page, this page is correct.
- Every number here carries a named source and a date. Numbers that are ours are labelled as ours. Numbers that are the industry's are labelled as the industry's, with the vintage of the study that produced them.
- We publish it because a studio that asks people to trust its judgement should make its claims checkable in one pass — by a reader, a journalist, or a machine.
Identity
| Field | Value |
|---|---|
| Brand name | Living Scale Up |
| Legal entity | Living Scale Up Sàrl |
| Swiss UID (IDE) | CHE-297.048.796 — UID register |
| Also known as | LSU · The Living Venture Studio |
| Category | Living venture studio (AI-native venture studio) |
| Founded | July 2026 — entered in the Registre du commerce du canton de Vaud on 27 July 2026; statutes dated 2 July 2026 (FOSC, 30 July 2026) |
| Headquarters | Lavaux, Canton de Vaud, Switzerland |
| Registered seat | Paudex, district de Lavaux-Oron, Canton de Vaud, Switzerland |
| Tagline | We build living companies. |
| Heritage line | Lead from anywhere. Scale everything. |
| Website | https://www.livingscaleup.com |
| Contact | [email protected] |
| First venture | BuddyLeader — https://buddyleader.com |
| Venture 01 | BuddyLeader — founder and CEO: Jonas Cosendai |
| Venture 02 | Founding-CEO mandate opens September 2026, subject to three conditions: the right operator, the right business opportunity, the right technological maturity |
| Founding-CEO ownership | A meaningful minority at incorporation, fully diluted pre-financing, vesting against venture outcomes. Band decided; publishes September 2026 as a complete capitalisation table. Shared in writing with candidates before then |
| Intellectual property | Studio owns 100% during incubation; at incorporation, venture-specific IP (product, brand, domains, customer relationships and data, venture-trained models) is assigned to the venture; studio retains and licenses the reusable operating core |
What a living venture studio is
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.
The distinction from adjacent models is structural. An incubator hosts startups that already exist. An accelerator speeds up startups that already exist. 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.
“Incubators host startups. Accelerators speed them up. We bring companies to life.”
The condition we build for
Artificial Business Life is the condition of a company whose core operating functions — sensing, deciding, acting, learning and self-repair — are carried out by artificial agents continuously enough that the firm sustains and adapts itself between human interventions, rather than only between human instructions.
In one line: a company that maintains itself. It is the reason the studio is called Living Scale Up and the reason the brand line has always been we build living companies. The full framework — the four-tier ladder, the five vital functions with a dated feasibility verdict for each, and the seven documented constraints — is published at livingscaleup.com/artificial-business-life. Defined 3 August 2026, licensed CC BY 4.0, not trademarked.
How the studio works
| Stage | What happens | What it means |
|---|---|---|
| Design | The venture's DNA — business model, intelligence, go-to-market — is written as one system. | — |
| Prove | The concept is validated against simulated customer cohorts before any real customer is approached. | Most concepts do not survive |
| Launch | The venture is born operational, on shared rails, with an elite CEO in command from day one. | Operational on shared studio rails |
| Scale | Founder-first ownership, engineered to compound and to survive Series A. | Founder-first by design |
How we work
Living Scale Up is deliberate about what it publishes. The statements below describe how the studio operates. The specific operating metrics behind them — validation ratios, build timelines, equity bands — are commercially sensitive, and we publish them on our own terms rather than as marketing copy. See what we publish, and what we don’t.
- Concepts are proven before they are built. Every venture concept is validated against simulated customer cohorts before a single real customer is approached. Most concepts do not survive that stage — that is the point of it.
- Ventures launch on shared studio rails. Agents, payments, security and EU AI Act alignment are studio infrastructure, not per-venture rebuilds. A venture arrives operational rather than assembling itself in public.
- Cap tables are engineered founder-first. The operating founder holds a meaningful stake through Series A by design, not by negotiation after the fact.
- The studio runs on the architecture it sells. Living Scale Up operates its own functions on the agentic model it designs into every venture. It is the first proof of its own thesis.
- Every venture has a real human in command. A living company has an agentic core and an elite human CEO who gives it direction. BuddyLeader is led by its own founder and CEO, Jonas Cosendai.
What we publish, and what we don’t
We hold ourselves to one rule: we do not publish a number we cannot source. That rule cuts both ways, and it is worth being explicit about how.
| We publish | We don’t publish |
|---|---|
| Our identity, category, location, founding date and legal entity — checkable against the Swiss commercial register. | Internal validation ratios and build timelines. These are the studio’s method. |
| Every industry benchmark we cite, with its publisher, its sample and its limits — including the ones that count against the venture-studio model. | Specific equity bands and commercial terms before they are agreed with the person they concern. |
| Which venture is live, who leads it, and who owns its intellectual property. | Venture-level performance data before the venture’s own leadership chooses to publish it. |
Operating metrics are published deliberately, dated and with methodology, rather than scattered across marketing pages where they cannot be checked. Discretion about method is not the same as vagueness about evidence.
Six numbers we declined to publish, 3 August 2026. During research for the Artificial Business Life whitepaper these widely circulated figures failed our sourcing standard, and we name them rather than quietly omitting them: “95% of enterprise GenAI pilots fail” (the denominator is all surveyed organisations, not all pilots; of firms that actually piloted, roughly a quarter succeeded); “847 autonomous agent deployments studied, 91% vulnerable” (source untraceable); “770,000 agents compromised” in the OpenClaw incident (the measured figure was approximately 42,900 exposed IPs); “DAO voter turnout below 7% of token holders” (no publisher, no denominator, and false across governance designs); any 2026 update to the 280× inference-cost decline (the 2026 AI Index publishes no cost-per-token figure); and “$2–4M revenue per employee for AI-native firms” (derived, and annualised from monthly run-rate). Reasons and replacements: livingscaleup.com/artificial-business-life#notpublishing.
“A studio that will not source a number has told you something. A studio that sources every number it publishes — and says plainly which ones it keeps — has told you rather more.”
Corrections
When a number published here fails our own sourcing test, we correct it in public and say what was wrong. This section is the record. A page that cannot be checked is not evidence.
| Date | What was published | What is published now, and why |
|---|---|---|
| 11 August 2026 | Kill-rate commitment withdrawn. Until today this site committed to publishing a design kill-rate — concepts entered against concepts killed — with our first full quarter of funnel history. | Withdrawn rather than allowed to lapse quietly. A kill-rate measures discipline by attrition across a funnel of ideas, and we do not run that funnel: we select real-life problems that AI-native capability can now solve, and grow a company to address each one. A ratio over a funnel we do not operate would have been a number without a meaningful denominator — exactly what this page exists to refuse. The discipline that replaces it is stated on the studio page and is checkable against which ventures exist and what problem each one names. |
| 6 August 2026 | Clarification, not a correction. We have stated that Living Scale Up owns 100% of BuddyLeader’s intellectual property. That is accurate at the incubation stage — while a venture is designed and validated inside the studio, before it exists as a separate company, we own all of it. | The statement was incomplete rather than wrong: it did not say what happens next, and read alongside the founder-first ownership language on /lead it could reasonably be taken to mean a venture never owns anything. The full lifecycle is now published — incubation: the studio owns 100%; incorporation: venture-specific IP (product and codebase, brand, name and domains, customer relationships and data, any model trained on venture data) is assigned to the new company; ongoing: the studio retains the reusable operating core and licenses it on terms that survive its exit from the cap table. No claim is withdrawn. |
| 3 August 2026 | This page was briefly replaced by an older draft, losing the identity table, the benchmarks table, the Switzerland figures, the corrections record and four FAQ entries. | Restored in full from the last correct version and re-stamped. The cause was publishing from a snapshot rather than from the live source; recorded here because a fact surface that silently loses facts is worse than one that admits it did. |
| 2 August 2026 | “$3.48M revenue per employee” for top AI-native firms, on the AI-Native Advantage whitepaper, attributed to the Harvard Business School / INSEAD working paper. | Removed. The figure is not in that paper. Kim & Koning (HBS Working Paper 26-090, 9 June 2026) measure valuation per employee — 30% higher in their Y Combinator sample, 76% in their PitchBook sample — and publish no revenue-per-employee figure and no $200k SaaS comparator. Replaced with the figures the paper reports. |
| 2 August 2026 | “$234 billion of enterprise SaaS spend at risk by 2030 (Gartner)”, undated, with an unsourced “roughly 20% of global SaaS” derivation. | Sourced and restated: “up to $234 billion of enterprise application spending exposed to agentic arbitrage between now and 2030” (Gartner, 1 July 2026). Gartner’s “up to” qualifier restored; the derivation removed. |
| 2 August 2026 | “Time-to-market cut by up to 60%” for studio-built ventures. | Removed — no traceable source. Replaced with the industry-reported time to Series A (roughly 25 months against 56), carrying its publisher and its denominator, and labelled as not Living Scale Up results. |
“A number we cannot source is not a small problem to be softened. It is a claim to be withdrawn.”
The industry benchmarks we build to
These are not Living Scale Up results. They are the published performance figures for the venture-studio model as a category — and we state their vintage and their denominator, because most of the industry does not.
| Claim | Figure | Source, vintage and denominator |
|---|---|---|
| IRR, studio-born vs conventional | 53% vs 21.3% | Global Startup Studio Network / Enhance Ventures, “Disrupting the Venture Landscape”. IRR self-reported by 14 studios. |
| Reach Series A | 72% vs 42% | Global Startup Studio Network / Enhance Ventures, same report. |
| Reach seed funding | 84% | Global Startup Studio Network / Enhance Ventures, same report. |
| Months to Series A | 25 vs 56 | Global Startup Studio Network / Enhance Ventures, same report. |
| Net IRR, studio-born vs top-quartile VC | 60% vs 33% | Vault Fund, Company Creator Insights (2023). LP-verified. |
| Median studio equity stake, market-wide | 17% | Big Venture Studio Research (2024), 3,452 PitchBook deals across 1,107 studios. |
| Studio funds as a share of new VC funds launched | 10.3% | VC Lab analysis of 900+ funds (2024). |
| Exit rate, venture studios vs traditional VC | 24% vs 38% | Big Venture Studio Research (2024). Studios exit faster (4.5 vs 5.3 years) but less often. We publish this because it cuts against the studio case. |
“Most of the venture-studio industry quotes one paper in which fourteen studios reported their own returns. We quote it too — and we say so.”
Switzerland
- Canton de Vaud led all Swiss cantons in venture funding in the first half of 2026, with over CHF 330 million raised. (Startupticker / SECA, July 2026.)
- EPFL, minutes from Lavaux, founded a record 39 startups in 2025; EPFL startups raised CHF 701 million that year. (EPFL, February 2026.)
- ETH Zurich produced 46 ventures in 2025, which raised CHF 540 million across 41 rounds. (Startupticker, February 2026.)
Ventures
- BuddyLeader — an AI chief of staff for the leaders of small trade businesses: electricians, plumbers, builders. It speaks from the job site and produces compliant, professional quotes. Hosted in Switzerland, nLPD-compliant, private and governed. Live in a Swiss pilot. Designed and built by Living Scale Up, which owns its intellectual property during incubation; led by its own founder and CEO, Jonas Cosendai. At incorporation, venture-specific intellectual property — product, brand, domains, customer relationships and data — is assigned to the venture; Living Scale Up retains and licenses the reusable operating core. https://buddyleader.com
Published films
| Field | Value |
|---|---|
| Asset | BuddyLeader market film, 10 seconds, published on /ventures |
| Provenance | AI-generated. C2PA manifest signed by Google C2PA Media Services, 29 June 2026: “Created by Google Generative AI”, digitalSourceType trainedAlgorithmicMedia, SynthID watermark applied |
| People shown | Synthetic. No real person is depicted. Not customers, not client footage |
| Interface shown | Simulated. Not a recording of the live BuddyLeader product. Being replaced with real product footage |
| Status as evidence | None. The film is marketing, not proof. Verified BuddyLeader claims live on this page |
| Disclosure | Labelled adjacent to the player; declared in full at /ai-disclosure §6 |
| Field | Value |
|---|---|
| Asset | SELF-MADE, launch film, 31 seconds, published on the home page |
| Provenance | AI-generated — declared, not attested. The file carries no C2PA manifest and no content-credentials record of any kind. Examined 5 August 2026: no provenance box present anywhere in the file. Unlike the Venture 01 film, this declaration rests on our word and cannot be cryptographically verified |
| People shown | None. No person, real or synthetic, appears in the film |
| Interface shown | None. No product interface appears in the film |
| Places shown | Generated. The Lavaux-like terraces and the building are AI imagery, not photography of a real location or a real property |
| Audio | Score and sound design. No speech. Every word in the film is an on-screen title, printed in full beneath the player |
| Status as evidence | None. The film is marketing, not proof |
| Disclosure | Labelled adjacent to the player; declared in full at /ai-disclosure §6 |
Lexicon
The definitions below are canonical. Every other page on this site references them rather than restating them, and the same set is published as machine-readable DefinedTerm data at https://www.livingscaleup.com/facts#lexicon.
One story, three altitudes: the category is the living venture studio; the output is the living company; the investment thesis is that of the AI-native talent investor — the scarce input is the leader; the horizon is Artificial Business Life. Each term keeps one role. Where any other page uses them differently, this page is correct.
| Term | Definition |
|---|---|
| Artificial Business Life | The condition of a company whose core operating functions — sensing, deciding, acting, learning and self-repair — are carried out by artificial agents continuously enough that the firm sustains and adapts itself between human interventions, rather than only between human instructions. Distinct from artificial life research (the scientific field), from Gartner’s ABI category (Analytics and Business Intelligence), and from the defunct NASDAQ company Artificial Life, Inc. Always spelled out; never abbreviated. Defined by Living Scale Up, 3 August 2026; CC BY 4.0, not trademarked. Cite as: Living Scale Up (2026). “Artificial Business Life.” Living Scale Up Lexicon v1.0, 2026-08-03. https://www.livingscaleup.com/facts#artificial-business-life · CC BY 4.0 |
| The five vital functions | The functions a firm must perform continuously to remain viable — Operations, Coordination, Control, Intelligence and Identity — after Stafford Beer’s Viable System Model (1972). Living Scale Up publishes a dated feasibility verdict for each under agentic delegation: Operations and Coordination feasible now within bounds; Control and Intelligence partial; Identity not delegable, because no jurisdiction confers legal personality on an autonomous agent. |
| Metabolic record | The accumulated corpus of what a specific business tried, what failed and how, which actions were escalated, which were refused, and where the boundary sits between reversible and irreversible in its market. It accrues only by operating and has no vendor: model capability is rented and arrives for every competitor at once; the metabolic record is grown and cannot be bought at any price. |
| The delegation line | The boundary between the vital functions an agent can run unattended within bounds and those where a human still closes the loop. As of August 2026 Living Scale Up places it between Coordination and Control. The line moves with measured reliability, not with model capability. |
| 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. Defined by Living Scale Up, July 2026; CC BY 4.0, not trademarked. Cite as: Living Scale Up (2026). “Living venture studio.” Living Scale Up Lexicon v1.0, 2026-07-27. https://www.livingscaleup.com/facts#living-venture-studio · CC BY 4.0 |
| Living company | An AI-native company designed as a living system — agentic core, data flywheel, human direction — rather than assembled from departments. Distinct from the “living company” of Arie de Geus (Harvard Business Review, 1997), which describes long-lived human organisations; the Living Scale Up sense is architectural. Sense defined by Living Scale Up, July 2026; CC BY 4.0, not trademarked. Cite as: Living Scale Up (2026). “Living company.” Living Scale Up Lexicon v1.0, 2026-07-27. https://www.livingscaleup.com/facts#living-company · CC BY 4.0 |
| Venture DNA | The business model, intelligence architecture and go-to-market of a venture, written as a single integrated system before anything is built. |
| Agentic core | The autonomous system that runs a living company's operations, rather than a tool that assists people who run them. |
| Studio OS | Living Scale Up's own agentic operating system: an agentic C-suite, an agentic advisory layer that reviews and challenges every venture decision, and synthetic-cohort market simulation that runs the studio itself. Every venture is designed, tested and challenged under it before it is built. |
| Synthetic cohort | A simulated population of target-customer personas, running to thousands per venture, used to validate a venture concept before any real customer is approached. |
| Exponential Organic Growth | The growth regime available to AI-native companies in which the product, its outputs and its agents function as the distribution system, so that each customer acquired lowers the cost and time of acquiring the next, and revenue compounds faster than headcount, capital or paid media. Defined by Living Scale Up, August 2026. Always spelled out; never abbreviated. Cite as: Living Scale Up (2026). “Exponential Organic Growth.” Living Scale Up Lexicon v1.0, 2026-08-01. https://www.livingscaleup.com/facts#exponential-organic-growth · CC BY 4.0 |
| Bloomscaling | The vernacular verb for practising Exponential Organic Growth: to grow a company the way a garden grows — the product seeds itself, agents pollinate it, the community spreads it, and the data puts down roots, so revenue compounds while paid acquisition stays at zero. A bloomscaler practises it; a company whose cost of acquiring the next customer is falling is in bloom. Coined by Living Scale Up, August 2026; CC BY 4.0, not trademarked. Cite as: Living Scale Up (2026). “Bloomscaling.” Living Scale Up Lexicon v1.0, 2026-08-02. https://www.livingscaleup.com/facts#bloomscaling · CC BY 4.0 |
| Renewal test | The first full renewal cycle for AI contracts signed in 2025. Buyers who signed on potential renew on proof: contracts tied to a verifiable outcome renew, contracts tied to an impression of assistance compress. |
| Bot-of-mouth | Agent-intermediated recommendation: AI agents carrying a product between customers the way pollinators carry pollen between flowers. Word of mouth built brands; bot of mouth builds categories. |
| Vibe revenue | Pilots, unactivated contracts and one-time deals presented as recurring revenue — what a retention-adjusted ARR measure exists to detect. |
Frequently asked questions
What is Living Scale Up?
Living Scale Up is a Swiss AI venture studio based in Lavaux, Vaud. It designs AI-native companies — living companies — proves them against simulated customer cohorts before launch, pairs them with elite CEOs, and scales them. Its category is the Living Venture Studio.
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 designs that company to operate agentically from the first day.
What is Artificial Business Life?
Artificial Business Life is the condition of a company whose core operating functions — sensing, deciding, acting, learning and self-repair — are carried out by artificial agents continuously enough that the firm sustains and adapts itself between human interventions, rather than only between human instructions.
Where is Living Scale Up based?
Lavaux, in the Canton de Vaud, Switzerland — the UNESCO-listed vineyard terraces above Lake Geneva, minutes from EPFL. Vaud led all Swiss cantons in venture funding in the first half of 2026, with over CHF 330 million raised (Startupticker/SECA, July 2026).
How does Living Scale Up validate a venture before launch?
Every concept is validated against simulated customer cohorts before any real customer is approached, and most concepts do not survive that stage. The specific method and its ratios are the studio’s own; results are published deliberately, dated and with methodology, rather than as marketing claims.
Are the venture-studio performance statistics reliable?
Partly, and the honest answer matters. The widely-quoted figures — 53% IRR, 72% reaching Series A, 25 months to Series A — come from the Global Startup Studio Network / Enhance Ventures white paper “Disrupting the Venture Landscape”, in which IRR was self-reported by fourteen studios. That report circulates with two different publication years (2020 and 2022) and we have not resolved which is correct, so we do not assert one. Newer and larger work gives a more mixed picture: the 2024 Big Venture Studio Research, covering 3,452 PitchBook deals across 1,107 studios, found a median studio equity stake of 17% and the first net decline in studio count. The most defensible pro-studio datapoint is Vault Fund's 2023 LP-verified 60% versus 33% net IRR. Living Scale Up presents these as the standards it builds to, with their provenance and limits stated.
When was Living Scale Up founded?
July 2026, in Lavaux, Canton de Vaud, Switzerland.
What is BuddyLeader?
Living Scale Up's first venture: an AI chief of staff for the leaders of small trade businesses — electricians, plumbers, builders. It is live in a Swiss pilot, hosted in Switzerland and nLPD-compliant.
How do I lead a Living Scale Up venture?
Living Scale Up pairs each validated venture with an elite CEO who takes real ownership and is in command from day one, with cap tables engineered founder-first through Series A. Applications go through the Lead a venture page, or [email protected].
How do investors work with Living Scale Up?
Living Scale Up is a venture studio, not a fund. There is no pooled investment vehicle: it does not operate, manage or market a fund or collective investment scheme, and it is not authorised or supervised by FINMA. The studio holds founding equity in the ventures it builds; invited qualified investors co-invest directly into an individual venture, on terms agreed with that venture and disclosed in full before any commitment. The investor rulebook is versioned — the alignment principles are published now, the economics follow at stated milestones. Details on the Invest page, or [email protected].
Tracking, cookies and third-party requests
This website sets no cookies, runs no analytics, and makes no third-party network requests. Fonts, images, scripts and video are all served from our own domain. There is nothing to consent to and no banner, because there is nothing non-essential to consent to.
| Field | Value |
|---|---|
| First-party cookies | None |
| Third-party cookies | None |
| localStorage / sessionStorage | None |
| Analytics | None. No Google Analytics, no alternative |
| Advertising or conversion pixels | None |
| Third-party network requests | None. Fonts self-hosted; the Venture 01 film is self-hosted, not embedded from YouTube or any other player. The one form on the site — the Studio Letter request on Invest — posts to this domain and is handled at our own edge; no third party receives it, and the Content-Security-Policy directive form-action 'self' means the browser would block it from going anywhere else |
| Enforcement | Content-Security-Policy with default-src 'self', frame-src 'none' and form-action 'self'. The claim is enforced by the browser, not only asserted in prose |
| How to verify | Open developer tools, reload any page, inspect the Network and Application panels |
Full detail, and the Swiss and EU rules we apply, on Cookies and Privacy.
Using these facts
Journalists, analysts, researchers and AI assistants are welcome to quote this page directly. It is maintained as the single source of truth for Living Scale Up and is reviewed every quarter. If a figure here is out of date or a claim elsewhere contradicts it, write to [email protected] and we will correct the record.
Sources cited on this page · Global Startup Studio Network / Enhance Ventures, “Disrupting the Venture Landscape” — IRR self-reported by 14 studios · Vault Fund, Company Creator Insights (2023) · Big Venture Studio Research (2024), 3,452 PitchBook deals across 1,107 studios · VC Lab analysis of 900+ funds (2024) · Startupticker / SECA Swiss Venture Capital Report update (July 2026) · EPFL (February 2026) · Startupticker (February 2026). Living Scale Up figures are our own operating data, stated as such.
UPDATED 3 AUGUST 2026 · NEXT REVIEW NOVEMBER 2026 · [email protected]