Venture capital
The financing model that built the tech economy
Description
In 2023, venture-capital firms managed roughly $3 trillion in assets globally and deployed more than $300 billion into startups. From essentially nothing in the 1950s, the industry has grown to a scale that meaningfully affects national economies. The companies that dominate contemporary technology Apple, Google, Amazon, Facebook, Tesla, Netflix, Uber, Airbnb were all venture-backed at critical stages. The iPhone, the search engine, the cloud, the social network, the ride-share, the streaming service are all, in some sense, products of the VC industry's willingness to fund extremely risky bets in hopes of extremely large returns.
This is not a natural feature of capitalism. For most of economic history, entrepreneurs raised money from family, banks lending against collateral, or wealthy individuals who knew them personally. The idea that a professional class of investors would deploy other people's money into early-stage companies with no revenue, no collateral, and a ninety percent failure rate while producing adequate returns is a specific institutional arrangement, built deliberately, much younger than most other features of modern finance. The VC model emerged in the US after World War II, matured in the 1970s and 1980s, and became the dominant form of early-stage tech financing in the 1990s.
Understanding how VC actually works is prerequisite to understanding the contemporary tech economy. The specific economics the power-law distribution of returns, the ten-year fund lifecycle, the LP-GP structure, the pressure toward exits shape how founders behave, how companies are built, and which ideas get resourced. Many of Silicon Valley's pathologies and successes are direct consequences of the VC financial model, not of deeper cultural forces. The model is one of the most consequential financial innovations of the twentieth century.
The question we're asking: what is venture capital, how does it actually work, and why has it produced the specific outcomes the technology industry has?
What we'll see: the institutional structure, the power-law returns that drive the industry, the specific effects on founders and companies, and the limits of the model.
Table of contents
01The institutional structure
A VC firm is structured as a series of funds, each with a defined lifecycle. Investors are called limited partners, or LPs pension funds, university endowments, sovereign wealth funds, insurance companies, family offices, wealthy individuals. The partners who run the firm and make investment decisions are general partners, or GPs. LPs commit capital to a fund for ten years. GPs deploy it into startups over the first three to five years, support the companies through growth, and harvest returns through exits (acquisitions or IPOs) over the remaining life of the fund. At year ten, the fund winds down and remaining positions are distributed.
GPs make money two ways. Management fees, typically two percent of committed capital per year, cover operating expenses. Carried interest, typically twenty percent of the fund's profits after LPs get their capital back plus a preferred return, is where the real money is. A GP at a top firm can earn tens of millions per fund through carry. The two-and-twenty structure was borrowed from hedge funds and has been remarkably stable in venture over four decades, despite occasional LP pressure to negotiate it down.
02The power law
The defining feature of VC returns is the power law. Returns are not distributed evenly across a portfolio. The vast majority of companies return little or nothing failing outright or producing returns so modest they barely offset fees. A small number return two or three times. A very small number one or two per fund return ten, fifty, or a hundred times. These outliers are where essentially all the fund's total return comes from. A fund without at least one ten-bagger typically misses its IRR target.
The power law has been studied empirically across vintages. The shape is consistent. In a portfolio of thirty companies, ten or fifteen return zero, ten return between zero and two times, four or five return two to ten times, and one or two return more than ten times. The one or two outliers produce more than half the fund's total return. Sequoia's investment in Google, Accel's in Facebook, Benchmark's in Uber have generated returns that substantially exceeded the total capital ever deployed by their funds. These are not average outcomes. They are the outcomes the entire industry is built to produce.
03The effects on founders and companies
The VC model has specific effects on how founders run their companies. Pressure toward massive outcomes pushes founders to prioritize growth over profitability early. Pressure toward exits pushes founders to go public or sell, not run indefinitely as a private business. Pressure to deploy capital pushes founders to hire quickly, spend aggressively, scale before ready. These pressures are not bad in all cases in the right market with the right company, they produce the outlier outcomes the model is built for. In the wrong cases, they produce overextended companies that burn through capital and collapse.
The dynamics of successive rounds shape behavior. A company raising Series A at $20M valuation has two to three years to show progress for a higher-priced Series B. Failing means a flat round, a down round, or no round any of which signals weakness. The round progression creates artificial deadlines shaping how companies allocate attention. Companies that fit the pattern can ride it to enormous outcomes. Companies that do not often die not because they are bad businesses but because they cannot hit the specific milestones the model requires.
04The limits of the model
The VC model has worked well for specific industries during a specific period, but its limits are now clearer. The first: the power law depends on enough outliers to carry the fund. In markets where outliers are harder to produce because incumbents capture dominant positions, or billion-dollar opportunities narrow returns have been declining. The 2021 deployment peak has been followed by a correction where many funds are struggling to produce the returns they promised LPs.
The second limit is domain-specific. The model works when companies scale rapidly with modest upfront investment — the classic software pattern. It works less well for capital-intensive industries: biotech (decade-long drug development, billions of dollars), hardware (manufacturing scale requires capital before revenue), climate tech (infrastructure before commercialization). Adaptations deep-tech funds, biotech-specialist firms, climate funds often produce lower returns than software-focused funds, which creates pressure against the adaptations. The industries most important to the next generation's challenges may not fit the standard VC model.
05Conclusion
Venture capital matters as a subject because it is one of the most consequential financial innovations of the twentieth century and because its specific mechanics shape the contemporary technology industry in ways that are often invisible from outside. Understanding the VC model the fund structure, the power law, the round progression, the exit pressure is prerequisite to understanding why Silicon Valley looks the way it does, why certain kinds of products get built and others do not, and why the industry produces the specific mix of extraordinary successes and public failures it produces. The model is not neutral infrastructure. It is a specific financial arrangement with specific incentives and specific consequences.