How venture returns work
Venture returns come from a handful of very large winners, and most companies in a portfolio lose money. This guide covers how that shapes portfolio size, manager selection, the wait for cash and what a good result looks like.
TL;DR
- A few companies carry the fund. In one dataset of 7,000 investments, the 6% that returned 10 times or more produced 60% of all returns.1
- Most investments lose money. Across 21,000 venture financings, 65% returned less than the money put in.2
- Portfolio size matters. A fund with too few companies can easily miss the one winner it needs.3
- The manager matters most. Among 2019-vintage funds on Carta, a top-decile fund was at 3.01 times paid-in capital at the end of 2025, the median 1.33 times and the bottom quartile 1.02 times, mostly unrealised.4
- It takes a long time. Funds run 10 years or more, and US companies have been a median 11 years old when they list since 2001.5
Why do a few companies return the whole venture fund?
Venture returns follow a power law: a handful of companies grow so large that their gains outweigh the rest of the portfolio combined.
Horsley Bridge, which invests in venture funds, looked at just over 7,000 investments made by its managers from 1985 to 2014. Around half returned less than the original investment. The 6% that returned at least 10 times produced 60% of all returns.1
In venture the outliers set the average, so the useful question about a fund is whether it holds one or two companies big enough to pay for the rest. Managers back companies that could become very large and keep reserves to follow on in the ones that work. How a venture fund works covers the mechanics.
How many startups in a venture portfolio lose money?
Most of them. Correlation Ventures studied more than 21,000 venture financings from 2004 to 2013 and found that 65% returned less than the capital invested. Only 4% returned 10 times or more.2
Not every loss is total; some companies are sold below cost or return part of the money on wind-up. Even a good fund will have mostly disappointing names, and the failures show up early, so a fund often looks worse in its first years than it ends up. Reading a fund report covers those early statements.
In an ESVCLP, losses on eligible investments are disregarded for tax, so they can't offset gains elsewhere, and gains on eligible investments held 12 months or more are exempt. The ESVCLP guide has the detail.
Why does the number of companies in a portfolio matter?
If a few companies make all the money, a portfolio needs enough of them to have a fair chance of holding a winner. With five, missing it is the likely result.
In 2019 AngelList analysed thousands of deals syndicated on its platform and concluded that, at seed, investors would raise their expected return by backing every credible deal. In its simulations over ten-year windows, fewer than 10% of investors beat that approach, even with some skill at picking.3 AngelList has an interest in that finding, and in September 2026 the same researcher qualified it: investors decide which startups get funded at all, so "Seed can support broad diversification, but it cannot be reduced to passive indexing."6
Spreading wider narrows the range of outcomes without raising the average, which mostly means fewer portfolios that lose money.
Diversification simulator
Run 1,000 random portfolios of 5, 15, 30 or 50 companies and see how often each size loses money.
Open the diversification simulator (opens in a new tab)The simulator draws each company's outcome from an assumed spread in which about two in three companies return less than their cost and about one in 25 returns ten times or more.2 On that spread, about two in five portfolios of 5 companies lose money, against under one in ten portfolios of 50. The very best results get a little rarer, because one big winner moves a 5-company portfolio more than a 50-company one.
Many seed funds hold dozens of companies for this reason. How much to allocate covers spreading commitments across funds and years.
Why does manager selection matter more in venture than in other asset classes?
Because the gap between good and poor venture managers is far wider than between listed equity managers.7 Among 2019-vintage funds on Carta, a top-decile fund had a TVPI of 3.01 times at the end of 2025, a top-quartile fund 1.9 times, the median 1.33 times and a bottom-quartile fund 1.02 times. In the 2017 vintage the top decile was at 4.08 times and the median 1.89 times.4
The best funds pull far ahead of the rest
The power law explains the spread. A manager who owns enough of one of the rare very large companies has a good fund; one who misses them all can't make it up on the rest. Access matters too, because the strongest founders usually choose their investors.
Most of a client's result is decided when the manager is chosen. How to vet a venture fund lists the questions worth asking, starting with realised returns.
How long does a venture investment take to pay back?
Usually a decade or more. A typical fund calls capital over three to five years, then spends the rest of a 10-year term, often extended, growing and selling companies.
The big winners usually take longest. In Jay Ritter's US data, companies were a median 8 years old at listing in the 1980s, 5 in the 1999 to 2000 boom, and 11 across 2001 to 2025; in 2024 the median was 14.5 A seed fund backing a company in its first year or two can hold it for most of that time.
Trade sales can come in year three or four, but they rarely make the fund. Stage and fund size covers why companies stay private longer. Early statements mostly show valuations, so judge a fund on distributions once it's old enough to have some.
What does a good venture fund return look like?
The cleanest yardstick is the net multiple: dollars back for each dollar paid in, after fees and carry. Among 2017-vintage funds on Carta, now the most mature, the median was at 1.89 times at the end of 2025, the top quartile 2.53 times and the top decile 4.08 times, still partly unrealised.4
| Measure | What it counts | When it's useful |
|---|---|---|
| TVPI (total value to paid-in) | Cash returned plus the current value of what the fund still holds, divided by money paid in | Throughout the fund's life, but the unrealised part is an estimate |
| DPI (distributions to paid-in) | Cash actually returned, divided by money paid in | Later in the fund's life, when it's the number that matters most |
| Net IRR | The annual rate of return after fees, taking timing into account | Comparing funds of similar age; early cash returns flatter it |
Fund reports usually show all three. A fund can carry a high TVPI for years before any of it turns into cash.
Keep gross and net apart. A manager's deal-by-deal results are gross; after a typical 2% annual fee and 20% carried interest, the investor's net multiple is noticeably lower. The diversification simulator shows gross multiples.
Compare a fund with others of the same vintage year. On Carta's end-2025 figures, 2.0 times would be top quartile for a 2019 fund but below the top quartile for a 2017 fund.4 Reading a fund report walks through the benchmarks, and Venture in a portfolio covers how these returns sit alongside a client's other assets.
Frequently asked questions
Venture returns are concentrated in a few very large winners. In Horsley Bridge data on about 7,000 investments, the 6% that returned 10 times or more produced 60% of all returns.
It depends on the definition, but most venture investments lose money. Correlation Ventures found 65% of more than 21,000 financings returned less than the money invested, and only 4% returned 10 times or more.
There's no fixed number. More companies make it less likely a portfolio misses the few winners, so a five-company portfolio loses money far more often than one of 30 or 50, all else equal.
The spread between managers is very wide. Among 2019-vintage funds on Carta at the end of 2025, a top-decile fund was at 3.01 times paid-in capital, the median 1.33 times and a bottom-quartile fund 1.02 times.
Funds typically run 10 years or more, with most cash coming back in the second half. US companies have been a median 11 years old at listing since 2001.
Judge the net multiple against funds of the same vintage. Among 2017-vintage funds on Carta, the median was 1.89 times and the top quartile 2.53 times at the end of 2025. Past results don't predict future ones.
TVPI counts cash returned plus the estimated value of what the fund still holds. DPI counts only cash paid back, so it matters more as a fund matures.
Get client-ready PDFs of every guide
- A clean PDF of any guide to send a client
- An email when the rules change
- Early access to new guides and tools
- Nothing else, and you can unsubscribe any time
Talk to us about Fund III
Tall Order is a Melbourne seed fund. Fund III is a registered ESVCLP for wholesale investors. Justine Lum, our Head of IR, works with advisers and their clients.
Sources
- Benedict Evans (Andreessen Horowitz), citing Horsley Bridge, In praise of failure: Horsley Bridge data on just over 7,000 venture investments, 1985 to 2014. ↩
- Seth Levine (Foundry Group), citing Correlation Ventures, Venture outcomes are even more skewed than you think (2014): 65% of 21,000+ financings, 2004 to 2013, returned less than 1x; 4% returned 10x or more. ↩
- AngelList, Startup Growth and Venture Returns (Abe Othman, December 2019): broad indexing at the seed stage. ↩
- Carta, Q4 2025 VC Fund Performance (19 March 2026): TVPI by percentile for the 2017 and 2019 vintages. ↩
- Jay R. Ritter, University of Florida, Median age of companies at IPO. ↩
- AngelList, The Impossibility of “Indexing” Seed (Abe Othman, 18 September 2026). ↩
- Cambridge Associates, Venture Capital Positively Disrupts Intergenerational Investing (2020): Figure 5, return dispersion of venture managers against global equity managers. ↩