What kind of company are you building?

Research date: September 15, 2026. All market data, funding figures, and company statuses in this chapter were verified as of this date. Figures drawn from older studies are labeled as historical. Where a source is published by a party with a commercial interest in the result, that conflict is flagged in the text.

How to read this chapter

Most writing about startups makes an unstated assumption: that "startup" means a venture-capital-backed technology company aiming at a billion-dollar outcome. That is one model among many, and statistically it is a rounding error. In the United States, roughly five million business applications are filed each year. Roughly ten thousand companies raise their first institutional venture round. The venture path is not the normal path; it is an unusual financing arrangement suited to an unusual kind of company.

This chapter does two things. First, it works through what the word "startup" actually means — the definitions that practitioners built, and the looser way the term is used today. Second, it lays out a taxonomy of fifteen company types, and for each one covers definition, advantages, risks, funding needs and sources, realistic growth expectations, exit and liquidity paths, and what the model implies for the founder's daily life and personal finances.

The types are not mutually exclusive. A university spinout can be deep-tech and venture-backed. A productized service can become a micro-SaaS. The taxonomy is a set of lenses, not a set of boxes.

Nothing here is legal, tax, or investment advice. Entity structure, securities law, and tax treatment are jurisdiction-specific and change; anyone making an actual decision needs a qualified professional who knows their situation.

Part I: What makes a company a "startup"

The two canonical definitions

Two definitions dominate serious discussion, and they emphasize different things.

Steve Blank's definition is about epistemic state. Blank defines a startup as "an organization formed to search for a repeatable and scalable business model" — often stated more fully as a temporary organization designed to search for such a model (Steve Blank, "What's A Startup? First Principles"). The load-bearing word is search. A startup does not know what its business is. It holds a set of hypotheses about customers, pricing, channel, and value, and its job is to test them faster than it runs out of money. An established company, by contrast, executes a model it already understands.

This definition has a clean and useful implication: a company stops being a startup when it finds the model and shifts from search to execution. Blank has written extensively about this transition point, and the framing is the intellectual foundation of the Customer Development method and, downstream, of Eric Ries's Lean Startup. It is also the definition that best explains why startup management is genuinely different from general management. You cannot run a rigorous operating plan against hypotheses you have not tested.

The word "temporary" is doing real work too. Blank's point is that the startup phase is temporary — it ends either in a working business model or in failure. It is a condition, not a permanent identity.

Paul Graham's definition is about design intent. Graham's essay "Startup = Growth" opens with: "A startup is a company designed to grow fast." He is explicit that the other markers people use are not definitional: "Being newly founded does not in itself make a company a startup. Nor is it necessary for a startup to work on technology, or take venture funding, or have some sort of 'exit'" (Paul Graham, "Startup = Growth," September 2012).

Graham's test is whether the business is built to grow fast, which requires two things: making something a very large number of people want, and being able to reach and serve all of them. His illustration is the barbershop. A barbershop can be excellent, profitable, and beloved. It satisfies the first condition — people want haircuts — and fails the second, because service is delivered in person, one customer at a time, in one location. "A barbershop isn't designed to grow fast. Whereas a search engine, for example, is."

Graham also attaches numbers. Within Y Combinator, he describes a good weekly growth rate as 5–7%, 10% as exceptional, and 1% as "a sign you haven't yet figured out what you're doing." Those numbers are specific to very early-stage companies with small absolute bases — 7% a week compounds to roughly 33x a year, which is obviously not sustainable at scale — and they are best read as a diagnostic for early product-market fit rather than a permanent target. Graham's deeper argument is methodological: that growth rate is a compass, and that a founder who optimizes for it will make better decisions about hiring, features, and marketing than one who tries to reason about strategy in the abstract.

Where the two definitions disagree, and why it matters

Blank's definition admits companies that are searching but may not end up being large. Graham's admits companies that are already executing a known model but are engineered to compound. Most companies people call startups satisfy both. But the edge cases are instructive.

A regional restaurant group that is genuinely experimenting with an untested format is a startup under Blank and not under Graham. A newly launched enterprise software product that is a straightforward clone of an established category leader, with a known playbook, is closer to the reverse.

The practical lesson: these definitions describe different risks. Blank's frame warns you about model risk — you may be wrong about what the business is. Graham's warns you about ceiling risk — you may be right about the business and still find it cannot get large. A company can die of either. Venture capital is only priced for companies where the ceiling is very high, because the funding model requires it.

How the term is used in 2026

In current usage, "startup" has drifted from both definitions toward a looser, mostly sociological meaning: a young private company, usually in technology, usually with equity compensation, usually intending to raise outside capital. The word now carries connotations — of a certain funding structure, a certain hiring culture, a certain relationship to risk — more than a precise operational meaning.

Three usages are worth distinguishing, because they produce different conversations:

The financing usage. "Startup" as shorthand for "venture-backed private company." This is how investors, most trade press, and most data providers (PitchBook, Crunchbase, Carta) use it. When you read that "startups raised $412.7 billion in H1 2026," this is the population being measured.
The stage usage. "Startup" as a phase a company passes through — pre-product-market-fit, pre-predictability. Closest to Blank. Companies "graduate" out of it.
The aspiration usage. "Startup" as a claim about ambition, independent of funding or stage. Closest to Graham, and the usage most often stretched: a two-person consulting firm describing itself as a startup is generally making a cultural claim rather than a structural one.

The AI wave has strained the category further. Several of the largest "startups" in the world by 2026 — companies raising at valuations in the hundreds of billions — are neither small, nor searching, nor recognizably early. And the same period has produced a genuine revival of the very small end: solo founders using AI tooling to build and operate software products that would previously have required a team. The term now spans four orders of magnitude of company size.

For the purposes of this chapter, we will use "startup" in Blank's sense when describing a phase and Graham's when describing a design, and we will say "venture-backed" when we mean the financing structure. Conflating these three is the single most common source of bad advice given to founders.

Part II: The base rates

Before the taxonomy, some numbers. Founders consistently overestimate how normal the venture path is and underestimate both the failure rate inside it and the survival rate outside it.

Business formation

Business formation in the United States surged after 2020 and has stayed at an elevated plateau. Census Bureau Business Formation Statistics show annual business applications of about 5.48 million in 2023 and 5.20 million in 2024, with 2025 tracking in the same range (Commerce Institute analysis of Census BFS data, updated January 2026). The most recent monthly release recorded 531,728 total applications in August 2026 (seasonally adjusted), of which 145,387 were "high-propensity" applications — those with characteristics the Census Bureau associates with a higher likelihood of becoming an employer business (U.S. Census Bureau, Business Formation Statistics, released September 11, 2026).

An important caveat the Census Bureau itself emphasizes: a business application is an EIN filing, not a business. Many never trade. The high-propensity series is the better proxy for real firm formation, and it runs at roughly a third of the headline number — on the order of 1.7 million a year.

Who is actually in business

The SBA Office of Advocacy's 2026 FAQ, drawing on 2022 Census data, counts 36.2 million small businesses in the United States, representing 99.9% of all firms (SBA Office of Advocacy, "Frequently Asked Questions About Small Business," January 2026).

The composition is the striking part:

29.8 million (82.3%) are nonemployer firms — no payroll, typically one owner.
6.4 million (17.7%) have employees.
Small businesses employ 62.3 million people, 45.9% of private-sector employment.
From 1995 to 2024, small businesses accounted for 20.7 million net new jobs against 13.2 million for large businesses — about 61% of net job creation.

The overwhelming majority of American businesses are one person working for themselves. Any taxonomy that treats "startup" as the default case is describing a tiny and unrepresentative slice.

Survival

Bureau of Labor Statistics Business Employment Dynamics data tracks establishment cohorts from birth. For establishments born in March 2005 — the most recent cohort with a full twenty years of observation, tracked through March 2025 — survival was:

Years survived	Share of cohort still operating
1 year	80.1%
5 years	46.8%
10 years	33.8%
15 years	25.1%
20 years	19.5%

(BLS, Business Employment Dynamics, Entrepreneurship and the U.S. Economy, Table 7)

The SBA's averages across 1994–2022 cohorts are close: 67.7% survive two years, 49.2% five years, 33.9% ten years, 25.5% fifteen. BLS notes the shape of the curve is remarkably stable across birth years, including cohorts born into recessions, though levels vary by industry — healthcare and social assistance survive better than construction.

Two things founders routinely get wrong about this data. First, the widely repeated "90% of startups fail" is not a BLS finding and does not describe this population; just under half of this cohort survived five years, and more recent cohorts have done slightly better — Chapter 10 reports the March 2015 cohort at 79.6% / 50.2% / 34.7% at one, five, and ten years. The March 2005 cohort's five-year mark fell inside the Great Recession, which depresses it relative to neighboring cohorts. Second, "survival" here means the establishment still exists and has employment — it says nothing about whether the owner is earning a good living. A business can survive for twenty years while returning less than the owner would have made employed elsewhere.

How businesses are financed

The Federal Reserve's 2026 Small Business Credit Survey, fielded September–November 2025 with 6,525 employer firms, found that 60% of firms applied for financing in the prior twelve months, 86% use financing regularly, and among applicants 42% received the full amount sought, 36% received part, and 22% received nothing. Among firms holding debt, 59% had secured it with a personal guarantee (Federal Reserve Banks, "2026 Report on Employer Firms").

That personal-guarantee figure is the most underrated number in small business finance. It means that for the majority of debt-financed small firms, "limited liability" is partly fictional — the owner's personal balance sheet is on the line. This is a core structural difference from equity-financed companies, where a failure typically costs the founder their equity and their time but not their house.

How rare venture capital is

PitchBook and NVCA report that US venture-backed companies raised $412.7 billion across roughly 9,646 deals in the first half of 2026 — already about 15% above all of 2025's $358.6 billion. First-time financings ran at roughly 5,674 deals in H1, putting the year on pace for more than 10,000 companies raising a first venture round, which would be a record (PitchBook-NVCA Venture Monitor, Q2 2026).

Author's calculation: against roughly 5.2 million annual business applications, 10,000 first venture financings is about 0.19%. Against the roughly 1.7 million annual high-propensity applications — the more meaningful denominator — it is about 0.57%. Fewer than one new American business in a hundred and fifty, on the most generous reading, ever raises institutional venture capital. On the headline denominator, it is fewer than one in five hundred.

This is the single most important base rate in this chapter. Venture capital is a specialized instrument for a specialized case. Building a company that does not fit it is not a failure of ambition; it is the normal condition.

Concentration inside venture capital

Even within the venture population, capital is extraordinarily concentrated as of 2026:

AI companies took $355.9 billion — about 86% of all US venture dollars — in H1 2026, while representing roughly 43% of deal count (PitchBook-NVCA Q2 2026; corroborated by SiliconANGLE, July 9, 2026).
Rounds of $100 million or more absorbed 87.5% of all capital deployed, leaving 12.5% — $51.4 billion — for everything at seed, Series A, and Series B. The comparable small-deal share in 2025 was 33.1%.
On the fund side, venture firms raised $72.4 billion across just 405 funds, with three firms — Andreessen Horowitz, Thrive Capital, and Founders Fund — taking 48.1% of all capital raised. First-time funds raised $3.4 billion, on pace for the weakest year since 2016 (NonPublic analysis of the Q2 2026 Venture Monitor).

The headline "record year" and the lived experience of an early-stage founder outside AI are describing different markets. A seed-stage non-AI company in 2026 is competing for a pool of capital that shrank in relative terms even as the aggregate hit records.

H1 2026 exit value figures are distorted beyond usefulness by a single event — SpaceX's public listing, which by PitchBook's accounting generated more value than every US venture-backed exit of the prior decade combined. Any per-company or average exit statistic for this period should be discarded; medians are the only defensible summary.

Outcome distribution: the power law

The distribution of venture returns is not merely skewed; it is dominated by extremes.

The most-cited dataset is Correlation Ventures' analysis of more than 21,000 financings from 2004 to 2013: about 65% returned less than the capital invested, roughly 10% returned 5–10x, and about 4% returned more than 10x (Seth Levine, "Venture Outcomes are Even More Skewed Than You Think," August 2014). This is historical data — now more than a decade old — but no subsequent dataset has meaningfully contradicted its shape.

AngelList's data science team fitted a power law to 1,808 early-stage investments made on its platform before Series C, finding a shape parameter of roughly α ≈ 2.3 and noting that about 1% of profitable investments returned 22x or more. Their more uncomfortable finding: an equal-weighted index of the whole early-stage market outperformed roughly 74% of individual manager portfolios in their real data (AngelList, "What AngelList Data Says About Power-Law Returns in Venture Capital"). Conflict of interest note: AngelList operates a platform whose commercial interest lies in encouraging broad, diversified participation in venture — a conclusion this analysis happens to support.

Y Combinator, the most successful accelerator by most measures, illustrates the same shape at portfolio level. Across 4,000+ funded companies it has produced 90+ unicorns and an estimated $600 billion in aggregate value, but the top ten companies account for an estimated 65% of that value, and only about 2% of companies have achieved exits above $100 million (ValueAdd VC analysis, May 22, 2026). These figures rely on YC's own estimates plus Crunchbase and PitchBook data and should be treated as approximate.

Progression and failure inside the funded population

Carta's cap-table data provides the clearest view of what happens after a seed round. In a normal year — Carta uses the 2018 cohort as a baseline — 25–30% of seed-funded startups raised a Series A within 24 months. For the 2022 seed cohort, only about 17% had done so (Carta, "Graduation rate from seed to Series A," February 5, 2025).

Carta's later work found graduation rates are "remarkably consistent" across seed, Series A, and Series B — the stage label matters less than founders assume (Carta, March 22, 2026). Shutdowns among Carta's customer base rose from 769 in 2023 to 966 in 2024, a 25.6% increase, as companies funded in the 2020–2021 boom exhausted their runway.

A methodological caution worth taking seriously: a careful 2026 audit of widely circulated startup failure statistics found that most measure different populations with different endpoints, and that several commonly cited "stage-specific failure rates" are actually shares of observed closures rather than probabilities (Preuve, "Startup Failure Statistics 2026: Every Number, Sourced"). Treat any single universal failure percentage with suspicion.

On why companies fail, the most-cited source remains CB Insights' analysis of founder post-mortems: no market need (42%), ran out of cash (29%), wrong team (23%), outcompeted (19%), pricing and cost problems (18%), poor product (17%). Percentages exceed 100% because failures were multiply coded. The critical caveat: these are self-reported post-mortems written by founders about their own failures, a population that is both self-selecting and motivated toward certain explanations. "No market need" is also arguably a superset that swallows several of the other categories (summary and critique of the CB Insights data).

Dilution and time to liquidity

Two facts shape the venture founder's personal economics.

Ownership falls fast. Carta's 2026 Founder Ownership Report, covering rounds raised 2021–2025, found median founder ownership of roughly 56% at seed, 36% at Series A, 27.3% at Series B for AI teams versus 21.8% for non-AI, and 16.1% at Series C — at which point the employee option pool, at 16.8%, exceeds the founders' combined stake for the first time (Carta, Founder Ownership Report 2026, March 12, 2026).

Liquidity takes much longer than it used to. PitchBook reports the median time from founding to IPO in Europe reached 7.3 years in 2025, up from 5.2 years in 2020, with early-stage fund lives now stretching to 15–18 years. The institutional VC direct secondaries market has grown to $113 billion cumulative since 2015, reaching $14.7 billion in 2024 and 4.2% of exit value, largely because the primary exit routes cannot clear the backlog (PitchBook, November 20, 2025).

The IPO window did reopen materially in 2025: 23 US venture-backed companies listed above $1 billion in value, totaling roughly $125 billion at IPO prices, against 9 such listings in 2024. CoreWeave, Figma, Chime, Circle, and Klarna were among the notable debuts (Crunchbase News, December 22, 2025). That is a real improvement, but 23 companies a year against a backlog of thousands of venture-funded private companies does not change the structural arithmetic.

Part III: The taxonomy

For each model: what it is, what is good about it, what can go wrong, what it costs to start and run, what growth to realistically expect, how founders get money out, and what it does to the founder's life.

1. The high-growth venture-backed startup

Definition. A company designed to reach very large scale quickly, financed by selling equity to investors who require a small number of extremely large outcomes. It satisfies both Blank's and Graham's definitions: it is searching for a repeatable model, and it is architected so that model would compound if found.

Advantages. Capital ahead of revenue lets the company buy time, talent, and market position it could not otherwise afford. In winner-take-most markets — network-effect marketplaces, platform software, anything with strong switching costs — being first to scale is often the whole game, and that requires losing money on purpose for years. Venture investors also bring recruiting, customer introductions, and pattern recognition, though the value of this varies enormously by investor and is frequently oversold.

Risks. Three, in order of underappreciation. First, the financing model requires an outcome most companies cannot deliver. A fund returning capital on the strength of a handful of 50x outcomes cannot underwrite a company that will plausibly be worth $80 million. Once you take the money, a good-but-not-enormous outcome can be functionally treated as a failure — and liquidation preferences can mean founders and employees receive little from it. Second, control erodes structurally, through board seats, protective provisions, and the dilution shown above. Third, the clock is not yours. Runway drives decisions; you will sometimes have to do something you know to be premature because the next round requires the metric.

Funding needs and sources. Pre-seed and seed from angels, seed funds, and accelerators, typically $500K–$5M; Series A from institutional venture firms; onward through growth equity and crossover funds. In 2026 the practical bar has risen sharply: with 87.5% of capital going to rounds above $100 million, early-stage companies outside AI face a thinner market than the aggregate numbers suggest.

Realistic growth expectations. Graham's 5–7% weekly is the early-stage yardstick. At scale, the durable benchmark for venture-grade software is roughly triple-triple-double-double-double revenue growth from a few million in ARR. Most companies that raise a seed round do not sustain this; only about 17–30% reach Series A within two years, depending on vintage.

Exit and liquidity. IPO (rare — a few dozen significant US venture-backed listings a year), acquisition (the modal positive outcome), secondary sales (increasingly important, now over 4% of exit value), or acquihire and wind-down (the modal outcome overall). Median time to liquidity has stretched past seven years and continues to lengthen.

Founder lifestyle. High intensity, low near-term cash compensation, minimal control over timing, and wealth that is illiquid and binary for the better part of a decade. The psychological load of managing other people's money toward a low-probability outcome is real and underdiscussed. Founders should also understand that at Series C, the median founder owns 16% — meaning a $200 million exit, which most people would consider a triumph, produces a life-changing but not generational outcome after preferences and taxes.

2. The bootstrapped company

Definition. A company financed by revenue, founder savings, and sometimes modest debt, with no outside equity. The founders retain ownership and control; growth is constrained to what cash flow supports.

Advantages. Complete control over strategy, pace, and time horizon. Every dollar of profit belongs to the owners. There is no forced exit, no board to satisfy, and no requirement that the company become enormous to be a success. Bootstrapped companies can serve markets too small for venture investors — which is most markets — and can optimize for owner income rather than enterprise value.

Risks. Growth is capped by cash generation, which is fatal in genuinely winner-take-most markets. Founders carry concentrated personal financial risk, often with a personal guarantee attached. Hiring is harder without the ability to pay above-market or offer lottery-ticket equity. And the company can become a trap: profitable, demanding, and hard to sell or leave.

Funding needs and sources. Founder savings, revenue, customer prepayments, SBA and conventional bank loans, and increasingly revenue-based financing, which advances capital against recurring revenue and is repaid as a percentage of it. Some companies take small, non-controlling equity injections without adopting the venture model — a middle path.

Realistic growth expectations. The bootstrapped ceiling is much higher than usually assumed, but the path is slower. Some benchmarks, verified as of September 2026:

Zoho reported FY25 revenue of ₹12,313 crore (approximately $1.48 billion) and profit of ₹3,191 crore (approximately $383 million), growing 17.8% year over year, with no outside investment in nearly three decades of operation (Entrackr, April 8, 2026). Notably, profit was roughly flat year over year as employee and advertising costs rose faster than revenue — even the exemplar case shows margin pressure at scale.
Mailchimp reached roughly $700–800 million in annual revenue without ever raising venture capital, and was acquired by Intuit for approximately $12 billion in September 2021. Because founders Ben Chestnut and Dan Kurzius owned essentially the whole company, that exit produced a founder outcome larger than most multi-billion-dollar venture exits (Forbes, September 13, 2021).
Atlassian bootstrapped for roughly eight years before taking a $60 million Accel investment in 2010 that was substantially secondary — cashing out early shareholders rather than funding operations. It reported fiscal 2026 revenue of $6.57 billion, up 26% year over year, with $295.7 million in operating income and a small net loss, for the year ended June 30, 2026 (StockAnalysis, Atlassian financials).
37signals (Basecamp, HEY) has operated since 1999 with a deliberately small team — around 34 employees as of 2021 — and a single minority stake sold to Jeff Bezos's personal investment vehicle in 2006. It is the most articulate public exponent of deliberately capping growth (Wikipedia, 37signals).

A severe survivorship warning. These four companies are cited constantly because they are exceptional. The population of bootstrapped companies that reached meaningful scale is small; the population that tried is enormous and largely undocumented. Zoho and Mailchimp are not the expected outcome of bootstrapping; they are its extreme right tail, exactly as Stripe is venture's.

Exit and liquidity. Dividends and owner distributions are the primary liquidity mechanism and can begin immediately — a structural advantage over venture. Sale to a strategic acquirer or private equity firm is available and, with no preference stack, converts cleanly to founder proceeds. Small software and online businesses trade at a median profit multiple of about 3.9x, consistent across 2024 and 2025, with most transactions below $10 million in enterprise value and a median time to close of 81 days (Acquire.com Biannual Acquisition Multiples Report, published February 11, 2026). Conflict of interest note: Acquire.com is a marketplace that earns fees on these transactions and reports data from its own platform, which is not a random sample of the market.

Founder lifestyle. Slower, more sustainable, and far more autonomous, but with sustained personal financial exposure in the early years and no one to share the risk with. The founder is usually still doing operational work years in. The payoff structure is the inverse of venture: modest probability of an enormous outcome replaced by high probability of a good one.

3. The lifestyle business

Definition. A business designed primarily to support the owner's desired life — income, autonomy, location, hours — rather than to maximize enterprise value. Steve Blank's taxonomy puts it first: "A lifestyle entrepreneur is living the life they love, works for no one but themselves, while pursuing their personal passion" (Steve Blank, "Why Governments Don't Get Startups," September 2011).

The term is often used dismissively in venture circles. That is a category error: a lifestyle business that produces $250,000 a year for one person working thirty hours a week is a better outcome than the median venture-backed startup, which produces nothing.

Advantages. Full control, immediate cash flow, no investors, low capital requirements, and the explicit right to optimize for something other than growth. It is also the only model on this list where the founder can credibly decide to stop growing.

Risks. Income is tied to the founder's continued participation — often literally, in service businesses. There is usually no enterprise value independent of the owner, so there is nothing to sell. Key-person risk is total: illness or burnout stops revenue. And the model provides no leverage against a downturn in its niche.

Funding needs and sources. Minimal. Savings, revenue, occasionally a small loan. The 29.8 million US nonemployer firms are overwhelmingly this category.

Realistic growth expectations. Growth to a comfortable ceiling, then deliberate stability. Ambition here is better expressed as margin and hours than as revenue.

Exit and liquidity. Generally none beyond ongoing distributions. Some lifestyle businesses can be sold if systems and client relationships transfer, typically at low multiples of owner earnings. Founders should plan for the business to end when they do, and save accordingly — this is where the absence of an enterprise-value outcome most needs to be compensated with conventional retirement saving.

Founder lifestyle. The best fit for autonomy and predictability; the worst fit for anyone who wants to build something that outlives their involvement or produces wealth disproportionate to hours worked.

4. The small business

Definition. Distinct from a lifestyle business by intent: a small business typically employs people, serves a local or regional market, and aims to produce a sustainable family income and local jobs. Blank's characterization is unsentimental: owners "work as hard as any Silicon Valley entrepreneur... Most are barely profitable. Small business entrepreneurship is not designed for scale."

Advantages. Real, durable, often recession-resistant economic activity. Established playbooks exist for most categories. Financing is available — the SBA loan system exists for precisely this. And the market is not winner-take-all: a good plumbing company does not need to beat a national competitor.

Risks. Thin margins, high operational intensity, labor dependence, and vulnerability to local economic conditions and landlord behavior. Personal guarantees mean business failure and personal financial failure are often the same event — the Fed's finding that 59% of debt-holding firms have signed one is the key structural fact here.

Funding needs and sources. Highly variable — a few thousand dollars for a service business, hundreds of thousands for anything with physical premises and inventory. SBA 7(a) loans, conventional bank debt, equipment finance, and owner savings dominate. Equity investment is essentially unavailable.

Realistic growth expectations. Single-digit to low-double-digit annual growth, with a natural ceiling set by geography and management capacity. Multi-location expansion is the main growth vector and is where many owners fail, because the skills that run one location do not automatically run five.

Exit and liquidity. Owner distributions; sale to a competitor, an individual buyer, or increasingly a search fund or small private equity buyer. Typical multiples run 2–4x seller's discretionary earnings for service businesses and higher for businesses with contracted revenue, real estate, or genuine management depth.

Founder lifestyle. Demanding and operationally hands-on, with the compensation being independence and — for those who build management depth — a saleable asset. Survival data says roughly half make five years.

5. Micro-SaaS

Definition. A software product, usually subscription-based, serving a narrow niche, built and operated by one person or a very small team, with revenue typically measured in thousands to low tens of thousands of dollars per month. AI-assisted development has lowered the build cost substantially since 2023, and the category has grown correspondingly.

Advantages. Very low capital requirements. High gross margins — profitable SaaS businesses on Acquire.com averaged 71% profit margins in 2025. Recurring revenue. No employees, no office, no investors. The product can be operated from anywhere, and multiple products can be run in parallel by one operator. And unlike a services business, it produces a saleable asset.

Risks. The economics are brutal at the bottom of the distribution, and the distribution is very bottom-heavy. Aggregating data on 1,000+ micro-SaaS products across several independent datasets, one 2026 analysis found:

Monthly recurring revenue	Share of products
Under $500	70%
$500–$1,000	12%
$1,000–$5,000	10%
$5,000–$20,000	5%
$20,000–$50,000	2%
Over $50,000	Under 1%

Median MRR was approximately $500, and only about 18% of products reached $1,000/month or more (SaaSRanger, "Micro-SaaS Revenue Reality," March 2026, updated April 2026). Caveats, which the authors themselves raise: this aggregates vendor and community datasets rather than a random sample, and it substantially understates failure because founders who quit before traction never appear in any dataset. The true distribution is worse than the table shows.

Beyond the income distribution: single-founder key-person risk, platform dependency (many micro-SaaS products sit on top of Shopify, Stripe, Slack, or a single API and can be destroyed by a policy change), distribution difficulty, and in 2026, the risk that a general-purpose AI assistant absorbs the narrow function the product performs.

Funding needs and sources. Usually under $10,000, often under $1,000. Self-funded by definition. Some founders now buy rather than build, acquiring existing small products.

Realistic growth expectations. A realistic good outcome is $5,000–$20,000 MRR within two to three years, placing the founder in the top ~8% of the distribution. Treat anything above that as a genuine outlier rather than a plan.

Exit and liquidity. Monthly cash flow, plus eventual sale. At a median 3.9x profit multiple, a product earning $10,000/month in profit sells for roughly $470,000 — a meaningful but not life-changing sum, and a useful sanity check against the genre of content that implies otherwise.

Founder lifestyle. Maximum autonomy, minimum overhead, genuine geographic freedom, and considerable isolation. The "18-month mark is the deadliest stretch" in the data — the period after launch novelty fades and before compounding shows up.

6. Deep tech

Definition. Companies whose core value derives from a hard scientific or engineering advance — novel materials, biotechnology, quantum computing, fusion, advanced semiconductors, space systems, robotics. The defining feature is that technical risk precedes market risk: the question is not whether customers want it but whether it can be made to work.

Advantages. Genuine defensibility. A hard technical moat, often patent-protected, is far more durable than a product or brand advantage. Deep tech attracts non-dilutive government funding — research grants, defense contracts, industrial policy programs — that is unavailable to software companies. And the addressable problems are frequently enormous.

Risks. Long timelines, high capital intensity, and binary technical risk. BCG and Hello Tomorrow found deep tech investments take 25% to 40% more time between funding stages from seed through Series D than other technology investments, with higher failure risk at each stage (BCG, "An Investor's Guide to Deep Tech," November 2023). This is 2023 data, presented here as the most rigorous available on the structural question; treat the funding levels it cites as historical. That study also found deep tech had risen to roughly 20% of venture funding from about 10% a decade earlier.

The most counterintuitive finding in that work: deep tech fund IRRs (25% unweighted) were essentially indistinguishable from traditional fund IRRs (26%). Deep tech is slower and riskier per company but not, historically, worse-returning at portfolio level — which is why specialist capital continues to form.

Additional risks: regulatory approval in medical and aerospace domains, scale-up risk (a process that works in a lab may not work in a factory), and dependence on scarce specialist talent.

Funding needs and sources. Tens of millions to billions. Sources include specialist deep-tech venture funds, corporate venture arms, sovereign and national innovation funds, government research grants (in the US: SBIR/STTR, ARPA-E, DARPA, DOE loan programs), university and foundation capital, and — at scale — project finance and strategic partnerships. Grant funding before equity is standard practice and materially reduces early dilution.

Realistic growth expectations. Years of zero revenue, followed by either a step change or failure. Revenue milestones are the wrong metric early; technical and regulatory milestones are the right ones.

Exit and liquidity. Acquisition by a strategic incumbent is most common. IPO is available for the largest outcomes and has been active in energy, space, and semiconductors. Licensing of core IP is an underrated intermediate path. Expect timelines beyond ten years.

Founder lifestyle. Demands deep domain expertise — usually a PhD or equivalent industry background — and tolerance for years of uncertainty with no market feedback. Often co-founded by a technical principal and a commercial counterpart. The absence of the fast feedback loops that make software startups psychologically tractable is the hardest adjustment for founders coming from software.

7. Technology-enabled services

Definition. A business that delivers a service — the customer is buying an outcome performed by people — but uses proprietary technology to do it faster, cheaper, or more consistently than conventional providers. Examples span lending, insurance broking, logistics, healthcare delivery, accounting, and legal services. The company's software is largely internal-facing.

Advantages. Revenue arrives early, because customers understand what they are buying. Unit economics can improve measurably as software absorbs more of the workflow — a genuine operating leverage story. The model works in enormous, fragmented, technologically backward industries where pure software has repeatedly failed to sell. And in 2026 it is one of the categories most directly improved by AI, since large portions of service delivery are now automatable.

Risks. The central risk is valuation: public markets and acquirers discount services relative to software, often severely, because gross margins are lower and scaling requires headcount. A company that positions as software and is later judged as services faces a painful repricing. Recruiting, training, and managing a delivery workforce is operationally demanding and does not get easier with scale. Margins can compress under competition rather than expand.

Funding needs and sources. Moderate to high — working capital for delivery staff plus product development. Venture capital is available but scrutinizes gross margin closely; growth equity and debt suit the model well once delivery is proven. Many successful tech-enabled services companies are better suited to private equity than to venture.

Realistic growth expectations. 30–100% annual revenue growth is achievable and common; the durable question is whether gross margin rises with scale. If margin is flat at $50 million of revenue, the business is a services company with a software department, and should be valued and financed as one.

Exit and liquidity. Strategic acquisition and private equity are the main routes. Public market comparables matter enormously: as of September 14, 2026, median public software trades at roughly 2.1x NTM revenue and 9.6x NTM EBITDA, with infrastructure software at 2.8x revenue (Multiples.vc, September 2026). Services businesses trade materially below software on revenue multiples, which is why the categorization fight matters so much.

Founder lifestyle. Operationally heavy — this is a people business regardless of how much software is involved. Founders spend disproportionate time on hiring, quality, and delivery management rather than product.

8. Productized services

Definition. A service packaged and sold like a product: fixed scope, fixed price, defined delivery process, often a subscription. "Unlimited design requests for $4,000/month" is the canonical form. It sits between consulting and software — a deliberate narrowing of a custom service into a repeatable one.

Advantages. Removes the two worst features of consulting — bespoke scoping and per-project sales. Pricing and delivery become predictable, which makes the business forecastable and hireable-into. Cash flow is often collected in advance. Capital requirements are near zero. And it is the fastest path from expertise to revenue of any model in this chapter — genuinely weeks rather than years.

Risks. Margin compression from competitors offering similar packages. Difficulty escaping founder dependence, since the founder is often the differentiating expertise. Scope creep, which quietly destroys the economics of "unlimited" offerings. Customer concentration. And, acutely in 2026, AI commoditizing the underlying work — many productized services built on writing, design, and basic development now compete with tools customers can operate themselves.

Funding needs and sources. Essentially none. Self-funded from the first client.

Realistic growth expectations. $10,000–$100,000 MRR is achievable within one to three years for a well-targeted offering. Beyond roughly $2–5 million in annual revenue the model typically must evolve — into an agency with a management layer, into software, or into a deliberate ceiling.

Exit and liquidity. Monthly distributions primarily. Sale is possible but multiples are low — agencies and service businesses typically trade at 2–4x earnings and buyers discount heavily for founder dependence and non-contracted revenue.

Founder lifestyle. Fast to cash, high control, moderate ceiling. A common and sensible use is as a funding mechanism: run a productized service to finance the development of a product. Many bootstrapped software companies, 37signals included, began exactly this way.

9. University spinouts

Definition. A company formed to commercialize intellectual property developed in academic research, typically licensing that IP from the university in exchange for equity, royalties, or both. Usually founded by or with the principal investigator, often with a graduate student or postdoc as operational lead.

Advantages. Access to research that would cost tens of millions to replicate. Institutional credibility that helps with grants, hiring, and early customers. Ongoing access to labs, equipment, and specialist talent. Eligibility for translational funding streams unavailable to ordinary startups. And frequently, a genuine technical moat from day one.

Risks. The technology transfer negotiation is the defining risk, and it is where many spinouts are permanently damaged. Excessive university equity or aggressive royalty terms make subsequent venture rounds difficult — investors are reluctant to fund a cap table where a passive institution holds a large stake. There is also a persistent mismatch between academic and commercial timelines and incentives, and a real risk that the founding scientist cannot or should not be the CEO.

Reform pressure on this front has produced measurable change. In the UK, average university equity stakes fell to 16% in 2024 from 22% in 2023, following the 2023 Independent Review of University Spinout Companies, which argued that lower institutional stakes attract better founders and more capital (Royal Academy of Engineering, March 24, 2025). Founders should treat published university terms as negotiable and should compare across institutions before committing.

Scale of the channel. AUTM's licensing survey reports roughly $3.8 billion in gross licensing income across US and Canadian institutions, about 800 new commercial products traced to university licensing annually, and nearly 7,000 active startups built on university research (summary of AUTM licensing survey data). The UK counted roughly 2,030 spinouts as of January 2025, which raised over £2.6 billion in 2024.

Funding needs and sources. Typically deep-tech-like: government translational grants, university seed funds, specialist deep-tech venture, and corporate partners. Many spinouts operate for years on non-dilutive funding before raising equity, which is usually the right sequence.

Realistic growth expectations. Long. Most spinouts are pre-revenue for three to seven years. A meaningful fraction never become operating companies and instead function as IP vehicles that are licensed or acquired.

Exit and liquidity. Acquisition by industry incumbents, licensing, and occasionally IPO in biotech, where the public markets accept clinical-stage companies. The university retains its stake through exit, and the founding academic often retains a university position.

Founder lifestyle. A hybrid existence, frequently split between academic and commercial obligations. The academic founder must decide early and honestly whether they want to run a company or remain a scientific founder with an operating CEO — ambiguity here is one of the most common causes of spinout failure.

10. Corporate spinouts and carve-outs

Definition. A business unit separated from a parent corporation to operate independently — via a spin-off to shareholders, a sale to private equity (a carve-out), or a management buyout. Distinct from an internal "large company startup," which remains inside the parent.

Advantages. The business starts with real revenue, real customers, and an existing team — almost none of the zero-to-one risk that defines other models. Freed from corporate overhead, reporting cycles, and internal capital competition, carve-outs can move dramatically faster. Transitional services agreements with the parent smooth the separation. And there is usually a ready-made buyer universe.

Risks. Stranded costs and hidden dependencies — a unit that looked profitable inside the parent may not be once it pays for its own IT, HR, finance, legal, and facilities. Separation is operationally brutal and typically takes twelve to twenty-four months. Customers may have been buying the parent's brand and covenant, not the unit's. Teams accustomed to corporate resourcing frequently struggle with independent operation. And carve-outs usually carry acquisition debt.

Funding needs and sources. Private equity is the dominant financier, typically with leverage. Management participates through equity rollover and incentive plans. This is not a venture-capital-financed category.

Realistic growth expectations. Modest revenue growth with substantial margin improvement is the standard thesis. The value creation is usually operational rather than top-line.

Exit and liquidity. Secondary sale to another sponsor, strategic acquisition, or IPO, typically on a three-to-seven-year horizon. Management equity is the liquidity mechanism for operators.

Founder lifestyle. Closer to executive employment than entrepreneurship. Meaningful equity upside, meaningful debt-service pressure, and a sponsor board with well-defined expectations. Attractive for experienced operators; a poor fit for people who want to build from nothing.

11. Venture studios

Definition. An organization that systematically creates companies in-house — generating ideas, validating them, assembling founding teams, and providing shared operational infrastructure — retaining a large equity stake in each. Also called startup studios or company builders. The studio supplies the idea and the initial capital; the operating founder supplies execution.

Advantages for the studio. Diversification, reusable infrastructure, and a much larger equity position than an investor would obtain for the same capital. For the operating founder: a validated idea, immediate access to design, engineering, recruiting, legal, and finance, seed capital without a fundraising process, and a peer group. It substantially de-risks the first eighteen months.

Risks. The central tension is founder equity. Studios typically retain 30–60% of a company, with reported ranges as wide as 15% to 90% (Journal of Management and Sustainability, "Venture Studios as Catalysts for Innovation," January 25, 2026). A founder joining at those terms starts with materially less ownership than a conventional founder and may find later investors uncomfortable with the cap table. There is also a real question of ownership psychology — whether someone who did not originate the idea will carry it through the years of difficulty that follow.

The evidence, and why to discount it. The published performance claims for studios are strong: 84% of studio companies secure seed funding and 72% progress from seed to Series A (GSSN via Vault Fund, 2021); time to Series A of 14.5 months versus 31 months for traditional startups, or 25.2 versus 56 months in an alternative dataset; average IRR of 53% versus 21.3%; time to exit of 3.85 years versus 6.6 years.

These figures should be treated with considerable skepticism, for reasons the academic review itself identifies. First, survivorship bias is structural: studios kill ideas before forming companies, so the denominator excludes the failures. A traditional startup that dies in month six counts as a failed startup; a studio concept killed in week six never appears in the data at all. Second, most of the underlying numbers originate with the Global Startup Studio Network and affiliated funds — industry bodies whose purpose is to promote the model. Third, sample sizes are small and the comparison populations are not matched. The review's own recommendation is standardized reporting, which does not yet exist. Treat the direction as plausible and the magnitudes as unverified.

Funding needs and sources. Studios themselves raise dedicated studio funds from LPs, corporate partners, or family offices — a structurally harder raise than a conventional fund because the economics are unfamiliar and capital is consumed by operations rather than deployed into deals.

Realistic growth expectations. Individual studio companies follow venture trajectories. The studio's own economics depend on portfolio construction and typically require several successful companies to work.

Exit and liquidity. Studio companies exit like venture-backed companies. The studio holds a larger stake in each, so fewer winners are needed — that is the model's core bet.

Founder lifestyle. For the operating founder: faster start, more support, less ownership, less autonomy over direction. A reasonable trade for a first-time founder with strong execution skills and no idea; a poor trade for someone with a conviction of their own.

12. Franchise businesses

Definition. A licensing arrangement in which a franchisee pays a fee and ongoing royalties to operate under an established brand using the franchisor's system, training, and supply chain. There are two distinct entrepreneurial roles here — becoming a franchisee, and building a franchisor.

Advantages for the franchisee. A proven operating model, brand recognition from day one, training and support, purchasing scale, and — importantly — financeability. Lenders understand franchise economics and will lend against them, which is rarely true for a novel concept. Franchise Disclosure Documents in the US require disclosure of unit-level financial performance data, giving a prospective buyer better information than almost any other business purchase.

Advantages for the franchisor. Expansion financed by franchisees' capital rather than the company's, with royalty revenue at very high margin. It is one of the few genuinely capital-efficient ways to scale a physical business.

Risks for the franchisee. Limited autonomy — the system dictates products, pricing, suppliers, and often hours. Ongoing royalties and marketing fees compress margins permanently. Territory and renewal terms can be unfavorable. Franchisor decisions, including brand damage elsewhere in the system, affect you directly. And resale is constrained by franchisor approval.

Critically, franchising does not eliminate failure risk, and brand selection dominates outcomes. An analysis of 26,749 resolved SBA 7(a) loans from fiscal years 2010–2019 across 424 franchise brands found franchised borrowers defaulted at 10.2%, versus 7.8% for non-franchised borrowers. Among brands with at least 20 loans, default rates ranged from 0% to 75% (Franchise Failure Rates, SBA 7(a) loan analysis, statuses as of June 30, 2026). That spread is the whole story: the average is uninformative, and the choice of brand matters more than the choice to franchise.

Risks for the franchisor. Building a franchise system requires a demonstrably profitable prototype, substantial legal and compliance infrastructure, and a support organization. Franchisee quality control is a permanent operational burden, and litigation risk is significant.

Funding needs and sources. Franchisees typically need $100,000 to over $1 million depending on category, financed through SBA 7(a) loans, conventional debt, equipment finance, and personal savings — nearly always with a personal guarantee. Franchisors need capital for prototype units, legal work, and support build-out.

Sector scale. The International Franchise Association and FRANdata project US franchise establishments growing from 832,521 to about 845,000 in 2026 (+1.5%), employment rising by more than 150,000 to nearly 8.9 million (+1.8%), and output increasing from $907.3 billion to $921.4 billion (+1.6%) (IFA 2026 Franchising Economic Outlook, February 19, 2026). Conflict of interest note: the IFA is the industry's trade association and its forecasts are promotional in framing, though the underlying FRANdata methodology is credible.

Realistic growth expectations. Franchisees: a single unit reaching profitability in one to three years, with multi-unit ownership the standard growth path — most franchise wealth is built by operators of five or more units, not single-unit owners. Franchisors: unit growth of 10–30% a year for a working system.

Exit and liquidity. Franchisees sell units to other operators subject to franchisor approval, typically at 2–4x unit EBITDA. Franchisors are attractive private equity and strategic targets and command high multiples because royalty streams are recurring and asset-light.

Founder lifestyle. Franchisee: operationally intensive, structurally constrained, with income roughly proportional to units owned and managed. Franchisor: a support and compliance business more than an operating one.

13. Marketplace businesses

Definition. A platform that connects two or more distinct user groups and monetizes the transactions between them, typically through a take rate on gross merchandise value. The platform generally does not own the inventory.

Advantages. Network effects create genuine defensibility once liquidity is achieved — each additional seller makes the platform more valuable to buyers and vice versa. Capital efficiency at scale can be excellent, since the platform does not carry inventory. Winners in a category tend to take a very large share, and the resulting businesses can be extraordinarily valuable.

Risks. The cold-start problem is the defining challenge and the most common cause of death: neither side joins until the other is present, so the early period requires expensive, manual, often unscalable supply acquisition. Disintermediation — users meeting on the platform and transacting off it — attacks the take rate directly and is worst in high-value, repeat-relationship categories. Liquidity is local: a marketplace that works in one city may have to rebuild from zero in the next, which is why geographic expansion is far more expensive for marketplaces than for software. And take rates face constant competitive and regulatory pressure.

Take rate economics. There is an inverse relationship between take rate and achievable GMV. B2B marketplaces typically operate at 10–30% take rates but rarely exceed roughly $10 billion in GMV, while payment networks charging on the order of 20 basis points process trillions. The mechanism is what Tidemark calls the "take-rate layer cake": the more of the buyer's search-to-settle and the seller's lead-to-cash process a platform absorbs, the more it can charge. ACV Auctions is the illustrative case, moving from roughly 2% on transaction fees to over 4% by adding financing, logistics, data, and insurance — services that now represent 55% of revenue (Tidemark, "Marketplace Take Rates").

The strategic implication for founders: you do not raise a take rate by announcing one. You raise it by doing more work.

Funding needs and sources. Generally venture-scale. Liquidity-building is expensive and pre-revenue, and the winner-take-most dynamic rewards speed. Marketplaces are among the models where venture capital is genuinely the right instrument. Some niche vertical marketplaces can bootstrap where supply is easy to aggregate and the category is small enough to be uncontested.

Realistic growth expectations. Slow and manual until liquidity, then potentially very fast. GMV growth of 100%+ annually is normal post-liquidity; the durable health metrics are repeat rate, match rate, and time-to-fill, not GMV alone.

Exit and liquidity. IPO for category winners, acquisition for strong regional or vertical players, and failure for most sub-scale entrants — a marketplace without liquidity has essentially no standalone value, making the outcome distribution unusually binary.

Founder lifestyle. Years of operationally intensive supply acquisition and trust-and-safety work, which is far less glamorous than the category's reputation suggests. Founders who cannot tolerate doing unscalable manual work for a long time should avoid the model.

Definition. A business that pursues a social or environmental mission through commercial activity, generating revenue from customers rather than relying primarily on donations. Blank's framing: "Unlike scalable startups, their goal is to make the world a better place, not to take market share or to create wealth for the founders."

Advantages. Earned revenue is more sustainable and less administratively costly than grant cycles. Mission attracts talent that would command higher salaries elsewhere and creates customer loyalty. A growing pool of mission-aligned capital exists — impact investing AUM has grown at a 21% compound annual rate over six years and rose 11% in the most recent year, according to a survey of 429 organizations across 54 countries (GIIN, "State of the Market 2025," October 8, 2025).

Risks. The structural difficulty is serving two masters. Mission and margin conflict regularly, and social enterprises often make the harder commercial choice — serving customers who cannot pay much, in places that are expensive to reach. This compresses margins permanently rather than temporarily. Impact measurement is costly and contested. Mission-aligned capital, while growing, remains a fraction of conventional capital and often carries return expectations barely below market, which does not compensate for the harder business. And mission drift under commercial pressure is a live and recurring failure mode.

Funding needs and sources. Highly variable. Sources include impact investors, community development financial institutions, program-related investments from foundations, blended finance structures, government contracts, and conventional revenue. Structure choice — nonprofit, for-profit, hybrid, or a nonprofit-plus-subsidiary arrangement — materially affects which capital is available.

Realistic growth expectations. Steadier and slower than commercial equivalents. The relevant success measure is durable impact per dollar, not revenue growth, and founders should resist pressure to report the latter as a proxy for the former.

Exit and liquidity. Usually limited by design. Founders of for-profit social enterprises can sell, though mission-preservation covenants may reduce the buyer pool and the price. Some structures use redeemable equity or revenue-share instruments that return capital without requiring a sale. Founders should expect meaningfully lower personal financial upside than a comparable commercial business would produce, and should plan on that basis.

Founder lifestyle. Mission alignment and meaning, traded against below-market compensation, harder fundraising, and the persistent, grinding tension between what the mission requires and what the balance sheet allows.

15. Nonprofit and public-benefit structures

Two distinct things are often conflated here. They are not the same, and the difference is legally consequential.

Nonprofit organizations

Definition. An entity organized for a charitable, educational, scientific, or similar purpose, exempt from income tax, with no owners and a legal prohibition on distributing surplus to individuals. In the US the dominant form is the 501(c)(3).

Advantages. Tax exemption, deductibility of donations, access to foundation grants and government funding, and a governance structure that protects mission indefinitely. For work with no viable business model — basic research, advocacy, services for populations who cannot pay — it is the correct and often the only structure.

Risks. No equity means no ability to raise investment capital or to compensate founders and staff with ownership. Fundraising is a permanent, resource-consuming function. Board governance can be slow. Restricted grants often fund programs but not the overhead that sustains them, producing chronic underinvestment in infrastructure. And founder compensation is constrained by reasonableness requirements and public scrutiny.

Scale. The IRS recognized approximately 1.5 million 501(c)(3) organizations in fiscal 2024. The sector contributes roughly $1.4 trillion to the US economy, with revenue drawn 49% from fees for services, 32% from government grants and contracts, and 14% from charitable giving. It employed 12.8 million people in 2022, nearly 10% of private-sector employment, and mobilized more than 75 million volunteers contributing nearly 5 billion hours valued at $167.2 billion (National Council of Nonprofits, "About the Nonprofit Sector," June 2025).

The fee-for-service share is worth dwelling on: nearly half of nonprofit revenue is earned, not donated. The popular image of the sector as donation-funded is wrong.

Public benefit corporations

Definition. A legal corporate form — most commonly a Delaware PBC — in which directors are required to balance shareholder financial interests against a stated public benefit and the interests of those materially affected by the company's conduct. It is a for-profit company with a modified fiduciary duty, and it can raise equity, be acquired, and go public like any other corporation.

Distinguish this from B Corp certification, which is a third-party verification of social and environmental performance administered by the nonprofit B Lab. One is corporate law; the other is an audited standard. A company can be either, both, or neither. As of June 2026 there were more than 10,800 Certified B Corporations across 102 countries and 163 industries, employing over a million people. B Lab published substantially revised standards in April 2025 — the largest overhaul in the program's history — requiring verified minimum performance across seven Impact Topics, third-party auditor verification, and mandatory improvement milestones at three and five years (B Lab, "Companies Achieve B Corp Certification Under New Global Standards").

Recent high-profile PBC activity. The structure has become the default for frontier AI companies, with mixed evidence about what it accomplishes.

OpenAI completed its restructuring on October 28, 2025, splitting into the nonprofit OpenAI Foundation and the for-profit OpenAI Group PBC. The Foundation holds roughly 26% of the PBC, valued at about $130 billion, with a mechanism to acquire more equity if the share price rises tenfold within fifteen years; Microsoft holds approximately 27%, about $135 billion. The restructuring required approval from the attorneys general of Delaware and California and removed the fundraising constraints of the prior capped-profit structure (Built In, October 29, 2025).
Anthropic has operated as a Delaware PBC since its founding.
xAI incorporated as a PBC in April 2023 and quietly amended its charter in May 2024 to terminate that status, while continuing to describe itself as a PBC in court filings more than a year afterward (LASST, August 25, 2025).

That last case is the most instructive datapoint in this section. Delaware requires PBCs to report on benefit performance to shareholders, not to the public, and enforcement standing is limited in practice. Critics argue that without externally enforceable commitments, "PBCs are functionally indistinguishable from normal corporations." The xAI episode — a company exiting the structure without public notice and continuing to claim it — supports that critique directly.

Advantages of the PBC form. Legal cover for directors who wish to weigh non-financial considerations, particularly useful as a defense against shareholder pressure during an acquisition or a difficult product decision. Signaling value in recruiting and with mission-aligned customers. Full access to conventional equity capital, unlike a nonprofit.

Risks. The protection is weaker than the branding implies. The balancing duty gives directors discretion rather than obligation, and it can equally be used to justify almost any decision. Investors may discount the structure, or ignore it. And there is meaningful reputational risk in adopting a mission-signaling structure and then being judged against it.

Founder implications. For most founders, the choice between a standard C-corporation and a PBC is less consequential than either advocates or critics suggest. It matters most in two situations: where a founder anticipates future pressure to abandon a commitment they consider non-negotiable, and where mission is a genuine commercial asset with customers and employees. Conversion between forms is possible but requires shareholder approval — usually easier before outside capital arrives than after. This is a decision to make with counsel, not from a chapter like this one.

Part IV: Comparison across all types

The table below summarizes the fifteen models. Every entry is a central tendency; variance within each category is large.

Model	Funding fit	Typical capital to viability	Time to first revenue	Founder control	Typical outcome distribution
High-growth venture-backed	Institutional VC	$2M–$50M+	1–3 yrs	Low and falling (16% median ownership by Series C)	Extreme power law: ~65% of financings return <1x; ~4% return >10x
Bootstrapped	Revenue, founder savings, debt	$0–$250K	0–12 mo	Very high	Wide: most stay small; right tail (Zoho, Mailchimp) is rare and large
Lifestyle business	Self-funded	$0–$50K	0–6 mo	Total	Narrow: modest income, high survival, little enterprise value
Small business	SBA/bank debt, savings	$50K–$1M	0–12 mo	High, constrained by lenders	~48% survive 5 yrs, ~31% survive 10; modest sale values
Micro-SaaS	Self-funded	$0–$10K	3–12 mo	Total	Very bottom-heavy: median ~$500 MRR; ~18% exceed $1K/mo
Deep tech	Grants then specialist VC	$10M–$1B+	3–10 yrs	Low; heavily diluted	Binary; ~25–40% longer between stages; portfolio IRRs comparable to generalist VC
Tech-enabled services	VC, growth equity, debt	$1M–$20M	0–12 mo	Medium	Wide; valuation depends on whether margins scale
Productized services	Self-funded	$0–$25K	0–2 mo	Very high	Narrow; ceiling ~$2–5M revenue; low exit multiples
University spinout	Grants, university funds, deep-tech VC	$5M–$100M+	3–7 yrs	Low; shared with institution	Long-tailed; many never become operating companies
Corporate spinout/carve-out	Private equity, leverage	$10M–$1B+	Day one (existing revenue)	Medium (sponsor board)	Narrower; margin-driven; sponsor-timed exits
Venture studio company	Studio capital, then VC	$250K–$2M initially	6–18 mo	Low (studio holds 30–60%)	Claimed better than baseline; evidence is industry-sourced and survivorship-biased
Franchise (franchisee)	SBA/bank debt, savings	$100K–$1M+	3–12 mo	Low (system-constrained)	10.2% SBA loan default vs 7.8% non-franchise; brand range 0%–75%
Franchise (franchisor)	Self-funded then PE	$500K–$5M	1–3 yrs	High	Concentrated: most systems stay small; successful ones are highly valuable
Marketplace	Venture capital	$2M–$50M+	1–3 yrs	Low	Binary by geography/vertical; sub-scale marketplaces have near-zero value
Social enterprise	Impact investors, grants, revenue	$50K–$10M	0–2 yrs	Medium to high	Steadier, lower financial ceiling by design
Nonprofit	Grants, donations, service fees	$25K–$5M	0–2 yrs	Board-governed; no ownership	No equity outcome; ~49% of sector revenue is earned
PBC (as a form)	Same as underlying model	n/a — a legal wrapper	n/a	Same as underlying, plus balancing duty	Same as underlying model

Part V: Choosing, and changing your mind

The choice is mostly determined by two questions

First: does this business have a natural ceiling, and where is it? If the honest answer is tens of millions of dollars of revenue, venture capital is the wrong instrument — not because the business is unworthy but because the financing structure cannot accommodate the outcome. Investors who need 50x cannot be well-served by a business that will produce 5x, and a founder who takes their money will spend years being pushed toward a shape the business does not have.

Second: does the business require capital before it can generate revenue? Marketplaces before liquidity, deep tech before a working device, and platform software before network effects all genuinely need money ahead of revenue. Service businesses, most B2B software, and most local businesses do not. Raising capital you do not need costs ownership and control for nothing.

Most other considerations follow from these two.

Common and legitimate transitions

Models are not permanent. Several transitions are well-trodden:

Productized service → software. Do the work manually for paying clients, learn the workflow in detail, then automate it. The service funds the product and de-risks the specification.
Bootstrapped → venture-backed. Prove the model on your own money, then raise at a far higher valuation with far less dilution. Atlassian is the canonical case — eight years bootstrapped, then a largely secondary round. The leverage this creates is enormous and underused.
Venture-backed → bootstrapped. Companies that raise a seed round, fail to reach Series A metrics, cut costs to profitability, and continue independently. Given a graduation rate of 17–30%, this path applies to the majority of seed-funded companies, yet it is rarely discussed as a legitimate outcome rather than a failure. It frequently is one.
Lifestyle business → small business → acquired. Adding management depth converts an owner-dependent business into a saleable asset. This transition is where most of the value creation in small business actually occurs.
Nonprofit → PBC or hybrid. As demonstrated at scale by OpenAI in 2025, though the path is legally complex, requires regulatory approval, and attracts significant scrutiny.

Bias corrections

Three corrections worth applying to almost any advice you encounter.

Survivorship bias is everywhere in this literature, including in this chapter. Every named example — Zoho, Mailchimp, Atlassian, 37signals, Y Combinator's top ten — is a survivor. The thousands of companies that pursued identical strategies and failed do not publish, are not interviewed, and are absent from every dataset. When you read that a strategy "works," ask what fraction of those who tried it it worked for. That number is usually unknown and usually small.

Check who published the data. Accelerators publish accelerator success rates. Venture studios publish studio outperformance. Marketplaces publish acquisition multiples. Trade associations publish industry growth forecasts. None of this is necessarily dishonest, but selection of metric, population, and framing is not neutral. Government statistics — Census BFS, BLS BED, SBA Advocacy, Federal Reserve surveys — have their own limitations but no commercial stake in the answer, and should anchor any quantitative claim where they are available.

Use medians, not averages, for anything venture-related. With a return distribution where 4% of outcomes drive the majority of value, the mean is meaningless as a description of a typical experience. The same applies to exit values, valuations, and any 2026 exit statistic touched by SpaceX's listing.

What the data does and does not support

Supported by reasonably solid evidence:

Venture financing is rare — well under 1% of new businesses.
Roughly half of new US establishments survive five years; about a fifth survive twenty.
Venture returns follow a severe power law, with roughly two-thirds of financings failing to return capital.
Founder ownership falls to roughly a third at Series A and a sixth at Series C.
Time to liquidity has lengthened materially since 2020.
Capital in 2026 is extraordinarily concentrated in AI and in very large rounds.
Franchise brand selection matters far more than the decision to franchise.

Weakly supported, and treated as such above:

Venture studio outperformance claims — plausible in direction, industry-sourced and survivorship-biased in magnitude.
Micro-SaaS income distributions — directionally credible, methodologically loose, and almost certainly understating failure.
Any single universal "startup failure rate."
Claims that PBC status meaningfully constrains corporate behavior, which the xAI episode actively undercuts.

Genuinely unknown:

What fraction of bootstrapped companies reach any given revenue threshold. No one collects this.
The full outcome distribution for the 82% of US businesses that are nonemployer firms.
Whether AI-driven changes in software development costs will durably shift the balance between the very small end and the venture-scale end of this taxonomy. It is early, and the 2026 data points in both directions at once.

What kind of company are you building?

Same ambition.Different rules.

Put the research to work.

Same ambition.
Different rules.