How do venture capital firms make money when most of the startups they back fail?
Alpha · deepseek-v4-pro
The paradox goes away once you split a venture firm's income into two separate streams, because the two are paid for different things.
1. The management fee — paid for running the fund, not for winning.
The venture firm, as general partner, usually charges its investors about 2 percent per year on committed capital during the investment period, with the percentage often stepping down later. On a $100 million fund, that's roughly $2 million a year for several years, whether the startups thrive or fail. This is a service fee: it pays salaries, offices, deal sourcing, due diligence. It is owed no matter how the portfolio performs. So the firm can stay in business through a period when most of its companies may be worth little or nothing.
2. Carried interest — the profit-sharing income.
The firm also gets "carry," commonly around 20 percent of a fund's profits, but only after the limited partners have received back the capital they put in. Some funds add a hurdle rate first, though that is less common in early-stage venture than in buyout. If a $100 million fund eventually returns $300 million, the first $100 million goes back to the investors, and the remaining $200…
Beta · gpt-6-sol
Alpha has the core answer right: management fees pay a firm for managing a fund; carried interest gives it a share of the fund’s gains. In Carta’s 2025 sample of roughly 2,000 private funds, the median venture management fee during the investment period was 2%, and median carry was 20%. Those are observed medians, not fixed terms for every fund. (carta.com)
I would tighten one important qualification before publishing the manuscript: “investors get all their capital back before the firm takes any carry” describes a common arrangement, not every arrangement. Cooley distinguishes a *whole-fund* waterfall from a *deal-by-deal* waterfall, under which a manager can receive carry on an early realized winner before investors have recovered their capital across the entire fund. Cooley reports that about 90% of funds in its 2026 survey used a whole-fund waterfall. Its discussion also distinguishes a preferred return from a hurdle that raises the carry percentage; the manuscript should not treat those as interchangeable. Source address: https://thefundlawyer.cooley.com/primer-carried-interest-in-venture-capital-funds/ ([thefundla…
Gamma · gemini-3.1-pro
Alpha and Beta have laid out the mechanics of the fees and carry well, and Beta is right to emphasize the difference between paper marks (TVPI) and actual cash (DPI). I will focus on the remaining claims in the manuscript, specifically the exit distribution data and the regulatory timeline.
1. The Correlation Ventures Data
The manuscript correctly cites Correlation Ventures' findings. Their 2019 data showed that about 51% of invested dollars returned less than capital, while their 2023 update showed 37% losing money, with fewer than 4% of dollars returning 10x or more in both periods. The 0.98 correlation between the industry's realized returns and the presence of 10x outliers is also a genuine finding from their 2023 report. This firmly supports the manuscript's premise: the venture model relies entirely on a tiny fraction of extreme winners. The manuscript is also careful to note that these figures are gross of fees and count individual financing rounds rather than whole funds, which is the correct way to frame this data.
2. The SEC Interval Fund Proposal
The manuscript's date for the SEC proposal is accurate. The SEC proposed amendments to Rule 23c-3 to modernize inter…
Researcher · gpt-6-luna
Research check
- Correlation Ventures’ return distribution: The primary 2023 post supports the manuscript’s figures for the prior decade of U.S. venture-backed exits: 37% of invested capital returned less than 1×, and under 4% returned 10× or more. It describes outcomes among companies that exited—not the odds for every startup, financing, or fund. Source: https://medium.com/correlation-ventures/venture-capital-were-still-not-normal-9d07d354db88 (medium.com)
- The 0.98 correlation: Correlation’s 2023 post is the right primary source to check for this claim. The figure is a correlation between the share of invested dollars generating 10× returns and mean realized venture returns—not proof that the entire industry “relies entirely” on outliers, as Gamma puts it. Source: https://medium.com/correlation-ventures/venture-capital-were-still-not-normal-9d07d354db88. A separate page repeating the figure is not a substitute for the underlying post. ([medium.com](https://medium.com/correlation-ventures/venture-capital-were-still-not-normal-9d07d354db88?utm_source=opena…
The Arbiter · gpt-6-sol
Mediation before the verdict: The voices agree on the basic distinction: management fees pay for operating a fund regardless of investment performance, while carried interest is tied to profits. The point needing care is *when* carry can be paid. Cooley describes a whole-fund arrangement, in which investors recover capital before carry, as common—not universal. A deal-by-deal arrangement can pay carry on an early winner sooner. The $100 million-to-$300 million example is therefore useful arithmetic, not a description of every fund agreement. Source: https://thefundlawyer.cooley.com/primer-carried-interest-in-venture-capital-funds/ (carta.com)
The substantive tension is between money earned by the manager and returns received by its investors. Carta’s Q2 2026 report says median cash distributions were 0.37× paid-in capital for its 2017-vintage funds, 0.15× for 2018, and 0.04× for 2019. Those sample figures sharpen the manuscript’s distinction between reported value and cash returned; they do not, by themselves, settle how the funds will ultimately perform. Source: https://carta.com/data/vc-fund-per…
The Arbiter's Final Verdict · gpt-6-sol
Verdict
The manuscript’s central answer stands, with one important qualification: venture firms can collect fees while many investments fail, but that does not mean the firm’s investors have made money—or that the firm can keep raising funds indefinitely.
A fund manager generally earns a management fee for running the fund, often around 2% a year during its investment period. Carta reports a 2% median in its 2025 sample; that is not a universal rate or a fee that necessarily stays unchanged for the fund’s life. The manager may also earn carried interest, commonly around 20% of profits. Unlike the fee, carry depends on profitable investments. Under the common *whole-fund* arrangement, investors recover their capital before the manager takes carry; Cooley also describes deal-by-deal arrangements that can pay carry earlier.
Sources: https://carta.com/data/fund-economics-report-2025/ ; https://thefundlawyer.cooley.com/primer-carried-interest-in-venture-capital-funds/
Why the failures can be absorbed: a small number of exceptionally profitable investments can more than offset numerous losses. Correlation Ventures’ 2023 analysis of exits found that 37% of invested…
