Unicorns, Chimaeras, & the Mathematics of Scale

The mythology was always the problem with the unicorn metaphor.  Unicorns are rare, beautiful, and imaginary.  The companies that venture capitalist Aileen Lee of Cowboy Ventures categorised under that term in her November 2013 TechCrunch article — “Welcome to the Unicorn Club: Learning from Billion-Dollar Startups” — are none of these things.  They are increasingly common, frequently ugly in their unit economics, and entirely real in their capacity to destroy investor capital.  Lee identified 39 unicorns in 2013.  By September 2021, CB Insights counted more than 800.  The variants have multiplied accordingly: a decacorn for companies valued above US$10 billion, a hectocorn for companies valued above US$100 billion.  The mythological creature that was chosen for its rarity has become the startup equivalent of a pigeon.

J. Skyler Fernandes of VU Venture Partners has proposed a more useful taxonomy for the tier above.  Companies worth US$1 trillion or more are chimaeras — the Greek mythological creature composed of multiple animals, typically a lion, a goat, and a snake.  Enormously powerful and actively consuming everything smaller than itself.  As of August 2021, six companies had crossed the US$1 trillion threshold: Apple, Microsoft, Amazon, Alphabet, Facebook, and Saudi Aramco.  All are platform businesses serving both consumers and enterprises simultaneously — with the exception of Saudi Aramco, which is a different category of beast entirely, powered by petroleum rather than software and sustained by a sovereign owner rather than market competition.

Tesla, at the time of writing, sat at approximately US$700 billion — the seventh largest company by market capitalisation, the next candidate for chimaera status, and a company of a distinctly different character from its predecessors.  Its revenue multiple of approximately 16x dwarfed the chimaera average of 7x, reflecting its earlier stage of maturity.  Its gross margin of 22% compared unfavourably with the chimaera average of 59% — a function of its physical manufacturing base rather than the digital-first architectures that make the existing chimaeras so margin-efficient.  Its EBITDA margin of approximately 14% sat below the chimaera average of approximately 41%, though Amazon — which reinvests aggressively at approximately 13% EBITDA margin — demonstrates that low margins and chimaera status are not mutually exclusive for companies whose reinvestment is producing structural advantage.

Berkshire Hathaway occupied an interesting eighth position at approximately US$645 billion in market capitalisation but crossed US$1 trillion in enterprise value — the calculation of market capitalisation plus debt minus cash — making it a chimaera by one measure and not by another.  This distinction between market capitalisation and enterprise value matters for investors who are comparing companies across different capital structures.  Berkshire Hathaway’s case illustrates why enterprise value is often the more meaningful number.

The Scale Problem Human Brains Cannot Process

The gap between a unicorn and a chimaera is not one of degree.  It is a factor of one thousand.  Most people understand this arithmetically and fail to internalise it practically.  Humans have a difficult time thinking at large scales.  Everything larger than a number with six or nine digits is simply described as “huge” — a cognitive limitation that likely reflects the fact that our ancestors never needed to contemplate trillion-dollar market opportunities to survive on the savannah.  This limitation has direct commercial consequences.  Founders who cannot think at a chimaera scale build unicorn businesses.  Investors who cannot think on a chimaera scale fund them.

The founders who pitch their total addressable market at US$10 billion and call it massive are not wrong about the number.  They are wrong about what it implies.  A US$10 billion TAM at 1% penetration produces US$100 million in revenue.  At a conservative 5x revenue multiple, that is a US$500 million exit — respectable, but not a chimaera.  Not even a decacorn.

To build a company worth US$1 trillion, you need revenues in the range of US$100 billion to US$1 trillion, depending on the revenue multiple the market applies.  Using the 1x to 10x revenue multiple range that characterises the majority of the top 100 companies by market capitalisation — the current chimaeras among them — a company must generate US$100 billion at a 10x multiple or US$1 trillion at a 1x multiple, with every combination in between representing the feasible range.

Facebook, the smallest chimaera by revenue at the time of the analysis, generated approximately US$104 billion at a 9x revenue multiple to reach its US$1 trillion valuation.  This is the lower bound empirical reference point for what chimaera revenue requirements actually look like.

The magnitude difference between a US$10 billion market size and a US$1 trillion market size is 100 times.  The magnitude difference between a US$10 billion market size and the largest known market on earth — real estate at approximately US$280 trillion — is 28,000 times.  These are not numbers that human intuition handles naturally.  They require explicit modelling to internalise.

The TAM Problem and Why Founders Lie to Themselves

The Total Addressable Market figure in most pitch decks is a work of creative fiction.  Founders cite research reports projecting market sizes in the tens or hundreds of billions — numbers produced by analysts making assumptions about entire theoretical categories rather than specific company customer bases and realistic revenue per customer.

The SAM — Serviceable Addressable Market — is the correct number.  It is the simple multiplication of actual potential customers by actual revenue per customer per year.  VU Venture Partners requires its investment team to present this equation explicitly in the investment committee: Market Size equals the number of potential customers multiplied by the average revenue per customer per year.  If the resulting number is below US$10 billion, the team must explain why a significantly higher-than-1% market penetration is achievable.

Two examples illustrate the practical difference.  Renting bridesmaid dresses: approximately 2.5 million US weddings annually, approximately five bridesmaids each, US$150 per rental — a SAM of US$1.875 billion.  At 1% penetration, US$18.75 million in revenue.  Not a unicorn opportunity.

Selling bed mattresses: approximately 132 million US households, approximately 2.5 beds per house, approximately US$1,000 per mattress replaced every eight years — a SAM of US$41.25 billion.  At 1% penetration, US$412.5 million in revenue.  A realistic path to a meaningful exit.

These two SAMs are 22 times different in size despite operating in the same consumer market.  The distinction is invisible if you describe both simply as “large consumer markets.”  It becomes decisive when you run the numbers.

The SOM — Serviceable Obtainable Market — is then the realistic revenue projection at exit, typically modelled as the revenue achievable at 1% penetration within five years: US$100 million-plus for a US$10 billion SAM.  Using a 1x to 10x revenue multiple, that produces a US$100 million to US$1 billion enterprise value at exit — the range that makes the venture model work for fund returns.

One notable exception: fintech companies typically trade at 10x to 20x revenues, reflecting the recurring revenue quality, high switching costs, and regulatory moat characteristics of financial services technology.  Most physical product companies trade at 1x to 3x revenues.  Pure recurring revenue software businesses trade at 4x to 10x.  The revenue multiple is not arbitrary — it reflects the quality of the revenue stream, the growth trajectory, and the gross margin profile of the underlying business.

The Mathematics of Venture Capital

The VC model is ruthlessly simple once stated explicitly.  A US$100 million venture fund needs to return US$300 million to be considered a strong performer.  With a typical portfolio construction of twenty to thirty companies, two or three exits must carry the entire fund.  The mathematics require each portfolio company to have the theoretical potential to return the entire fund, meaning the fund needs companies capable of US$1 billion-plus exits.

This is why VCs are structurally uninterested in good businesses with modest markets.  A company addressing a US$500 million market that executes perfectly generates insufficient exit value to matter to a fund, regardless of how well the business performs on its own terms.  The dilution mechanics compound the challenge.  Each financing round dilutes existing shareholders by approximately 20%, including new investor allocation and ESOP pool increases.  After three financing rounds — each at 20% dilution — an investor retains approximately 51.2% of their original ownership: 10% initial stake becomes 10% × 0.8 × 0.8 × 0.8 = 5.12%.  The initial 10x to 100x gross return potential becomes 5x to 50x after three rounds of dilution.

Early-stage VCs target a 10x-plus return, inclusive of dilution.  A company requiring four financing rounds to reach exit reduces a theoretical 100x return to approximately 40x — still excellent, but requiring a US$400 million exit on a US$10 million initial investment.  If the company exits at US$100 million, the return is approximately 4x — below the threshold that justifies the risk of early-stage investing when real estate consistently delivers 2x to 3x returns over five to seven years at materially lower risk.  This arithmetic is why the north star for founders and investors has shifted.  A unicorn exit at US$1 billion is the baseline aspiration.  The real prize is the US$10 billion-plus exit that returns an entire fund from a single position.

Bear Flag Robotics demonstrated the best-case scenario: VU Venture Partners invested at US$27 million and exited at US$250 million fourteen months later — approximately 9x gross with no dilution.  This is extraordinary in its speed and cleanliness.  It is also exceptional.  The systematic replication of this outcome is the aspiration, not the expectation, of the venture model.  VU Venture Partners also achieved three additional exits, yielding approximately 5%, 17%, and 650% IRR, respectively, within roughly twelve months of initial investment.  A win is a win when the alternative is a zero, and in venture capital, the zeros are the statistically normal outcome for most portfolio companies.

The Chimaera Characteristics

What distinguishes the six current chimaeras from the next tier of large companies is not merely scale.  It is a specific combination of structural characteristics that Fernandes identifies across the category.  They are platform plays.  They sell more than one product across multiple industries.  Their customers include both consumers and enterprises — and this dual-customer base is not incidental.  It is mathematically necessary.  With approximately 7.9 billion humans on earth and a finite number of enterprises, a company serving only consumers or only enterprises hits a structural ceiling that makes a US$1 trillion valuation almost impossible.

They operate across both physical and digital dimensions.  Amazon’s physical logistics infrastructure and Whole Foods grocery network sit alongside AWS cloud computing and Prime Video streaming.  Apple’s physical device hardware sits alongside the App Store, Apple Pay, Apple Music, and iCloud software and services.  The purely digital company cannot reach chimaera scale because the revenue per customer from digital-only services is insufficient without the volume that physical products generate.

They are prolific acquirers.  Google acquired YouTube for US$1.65 billion in 2006 — when YouTube was losing money — and DeepMind for approximately US$500 million in 2014, and dozens of other companies that have been absorbed with varying degrees of preservation of the original entity’s identity.  Facebook acquired Instagram for US$1 billion in 2012 when it had thirteen employees and no revenue.  At the chimaera scale, a US$1 billion acquisition represents 0.1% of enterprise value — the equivalent of a substantial corporation spending US$1,000 to acquire a meaningful competitive asset.  The risk-adjusted logic of absorbing potential competitors at 0.1% of enterprise value is overwhelming.

The acquisition strategy follows two patterns.  The acqui-hire model absorbs talent and terminates the original product, leaving minimal organisational trace.  The integrated subsidiary model preserves the acquired entity’s operational independence while connecting it to the platform’s infrastructure, data, and distribution.  Instagram, YouTube, and LinkedIn all followed variations of the latter.  The choice between patterns depends on whether the value sits in the people or in the product — and chimaeras are sophisticated enough to distinguish between the two before writing the cheque.

The Outliers in the Next 100

Among the companies below chimaera status but within the top 100 by market capitalisation, several valuation anomalies deserve examination.  Shopify traded at approximately 47x revenues — the largest outlier in the top 100.  This reflects the market’s conviction that Shopify’s recurring revenue model, high switching costs, and explosive growth trajectory justified a multiple that conventional revenue multiple frameworks could not contain.  Six companies — NVIDIA, Prosus, Visa, Moderna, Mastercard, and Adobe — traded between 20x and 24x revenues.  The next ten traded between 11x and 19x.  The vast majority — 80%-plus of the top 100 — traded between 1x and 10x revenues.

The distribution is not random.  It reflects the quality of revenue.  One-time product sale companies trade at 1x to 3x.  Recurring revenue businesses trade at 4x to 10x.  High-growth recurring revenue businesses with strong gross margins and expanding customer relationships trade at 10x to 24x.  Companies in genuinely exceptional growth phases with defensible network effects trade above 24x — and the market extracts a severe penalty when the growth fails to materialise at the implied rate.

The Future Chimaeras

The next generation of US$1 trillion companies will likely look different from the current six.  The current chimaeras are predominantly digital-first — companies of bits rather than atoms.  Future candidates may come from physical domains that the current digital chimaeras have not penetrated.

Nuclear fusion energy companies — if the engineering challenges are resolved — would address a US$10 trillion-plus global energy market.  Asteroid mining companies would access resource deposits whose combined value dwarfs terrestrial mining markets.  Space real estate and manufacturing companies would operate in a market with no physical constraint on the supply of building sites.  Biotechnology companies that address the entire addressable market of human disease would serve a customer base of every living human being.

Tesla’s candidacy for chimaera status — at approximately US$700 billion in mid-2021 and subsequently crossing US$1 trillion before retreating — demonstrates that physical manufacturing can reach chimaera scale.  Its path was through the intersection of hardware, software, energy, and financial services: selling cars that are also software platforms, generating regulatory credits that fund competitors’ compliance, and building energy storage infrastructure alongside vehicle manufacturing.

The pattern of diversification that produces chimaeras is consistent: companies that refuse to be bounded by their original product category, that expand into adjacent and non-adjacent markets simultaneously, that serve both consumers and enterprises, and that build platform infrastructure that others depend on rather than simply products that others purchase.

The Regulatory Question

The one variable that the chimaera taxonomy does not fully account for is antitrust.  Extreme concentrations of market power have historically attracted regulatory responses in democratic societies — the Standard Oil breakup of 1911, the AT&T breakup of 1984, and the Microsoft antitrust proceedings of the late 1990s all represent governmental interventions in markets that had consolidated to the point of structural concern.  The European Union’s Digital Markets Act — enacted in 2022 and progressively enforced through 2024 and 2025 — is the most aggressive recent attempt to constrain platform power without breaking up the companies themselves, imposing interoperability, data portability, and self-preferencing restrictions on the major platforms.

Whether the current chimaeras face structural breakup or continued regulatory constraint is the defining commercial and political question of the next decade.  A chimaera broken into its component parts — the lion separated from the goat separated from the snake — would produce multiple large companies from a single enormous one.  Whether the sum of the parts exceeds the whole depends on the degree to which the chimaera’s value is generated by platform integration rather than individual business unit performance.

Google’s advertising business, cloud business, YouTube, and hardware divisions each have standalone value.  Whether their combined value under Google’s integrated platform exceeds their sum as independent entities is the question that determines whether antitrust action creates or destroys shareholder value.  The historical precedent of AT&T — whose breakup created the Baby Bells that collectively exceeded the parent’s pre-breakup value — suggests the outcome is not predetermined.

The Conclusion

The chimaera metaphor is precise in one respect that Fernandes does not dwell on: a chimaera is classically defined as a thing hoped or wished for but in fact illusory or impossible to achieve.  The US$1 billion unicorn was once considered illusory.  It is now commonplace.  The US$1 trillion chimaera was once considered impossible.  Six of them exist.  The psychological marker that Fernandes identifies — the proof that what seemed impossible is now possible — applies at every order of magnitude.

The north star for founders and investors is no longer unicorn status.  It is a chimaera status — building a company that consumes unicorns rather than aspiring to become one.  That aspiration requires a fundamentally different approach to market selection, product strategy, customer base construction, and organisational ambition.

The VC is a unicorn hunter.  The founder who internalises the chimaera framework is building a different kind of animal entirely.  Greek mythology did not provide a creature that eats chimaeras.  That category remains to be named.  The founders building it are probably in a garage somewhere right now, pitching a market size that sounds implausible and a valuation that sounds impossible.  They are probably right about both.

---

Terence Nunis | Executive Chairman, Equinox Zenith | Author, The 1% Playbook: The Billionaire Cheat Code