Consistent with venture fund performance over the past 20 years, it will continue to be difficult for funds to generate a 3x return to investors within 10 years of their commitments, says Peter Walker, head of insights at Carta, on whose platform most VC-backed companies maintain their cap tables. But Walker is more optimistic about venture fund performance than he was two years ago.
“It’s only the top 20 percent, and in some cases only the top 10 percent, that are going to see those returns,” Walker tells Venture Capital Journal. “However, and we’re starting to see this with some of the massive AI companies in these very young funds, when you hit a winner, I think the winners are going to be even bigger than they’ve ever been before. If you’re a fund that started in the last two years and own a significant slice of Cursor or another one of these AI names, you’ve basically already returned your fund and more so.”
Only the top-tier venture capital funds with 2017 and 2018 vintages are positioned to deliver a return of at least 3x, according to Carta’s fund performance report for Q3 2025, released in December.
The report’s findings are based on data from 2,835 venture funds with vintages from 2017 through 2025. While most of the funds are smaller than $100 million in size, more than 300 are larger than $100 million. Altogether, the $100 million-plus funds claim to have roughly $61.9 billion in capital commitments, or 52 percent of a total of $118.3 billion in commitments across the full sample of all funds.
For 2017 vintage funds, the current 90th percentile for net TVPI (total value to paid-in capital) is 3.52x, while only 10 percent of funds with a 2018 vintage achieved a return of 3.07x or better. TVPI expresses the relationship between a fund’s total present value –including both realized and unrealized gains – and the amount of capital that LPs paid in.
The vast majority of funds raised in 2017 and 2018 are on pace for substantially lower returns. For the 2017 vintage, the median net TVPI through the third quarter is 1.76x, with just 25 percent of funds achieving a return of 2.27x or better. The median fund with a 2018 vintage has a net TVPI of just 1.38x, with a return of 1.97x or better for only 25 percent of funds, the report states.
As for IRR, “the median for 2021, 2022 and 2023 fund vintages all continue to lag well behind the performance of slightly older fund vintages at the same point in time,” the report notes. “After 15 quarters of management… the median IRR for the 2021 vintage sits at 0.2 percent. At this same signpost in time, the median IRR for the 2020 vintage was 4.3 percent. For the 2017 vintage, it was 26 percent.”
Due to a chilly exit environment, DPI, which measures realized returns distributed back to LPs, continues to be a rarity for most funds.
“At least half of all funds in each vintage from 2019 on have a DPI of zero, meaning they’ve yet to begin generating any net DPI,” the report states. “The 2017 vintage stands out as a clear exception. The median DPI for 2017 funds sits at 0.28x, which means the median fund in this vintage has distributed returns equivalent to about one-quarter of the initial capital that LPs paid in. The 75th percentile for DPI in the 2017 vintage rises to 0.51x, and the 90th percentile to 1.11x – meaning that these top-decile funds have begun to generate net-positive value for their LPs.”
Will 2021 be ‘worst’ vintage?
The portion of funds with a 2020 vintage that have begun to generate DPI increased from 28 percent to 42 percent over the past year. A greater percentage of funds with a 2021 vintage also issued distributions for the first time over the past 12 months, rising from 14 percent to 25 percent year-on-year.
“It’s still very early on in a lot of these funds’ lifecycles, but I think 2021 vintage funds will be some of worst-performing venture funds we’ve ever seen. It will be a difficult year to outperform in,” says Walker. While he thinks diversification is very important, the 2021 vintage is “such a clear drop-off from the years that came before it, and the years that came after it have also been better. So I’m a little worried about what’s going to happen to many of those funds and many of those fund managers over time.”
Part of the problem is the high valuations that vintage 2021 funds paid for investments. Another factor is that early 2022 is when a whole decade of low-interest rates got reversed, he notes. “The cherry on top of that is that the ChatGPT AI boom started in late 2022. So in all senses, it was investing just before the drop-off. I know many founders and many investors who are now basically entirely focused on AI-native companies. Any investments made in 2021 are by definition not AI-native, so it’s a struggling time period.”
Funds of more recent vintages currently sit well below the median returns generated by vintages prior to 2019. To a large extent, that’s to be expected. Nearly every venture fund raised since 2017 is still being actively managed, with plenty of time in the future for gains to accumulate. This is particularly true as venture-backed companies continue to stay private for longer, extending the projected lifespan of many VC vehicles.
Fewer LPs
Across all fund sizes, today’s new venture fund typically has fewer LPs than funds raised earlier this decade. For instance, the median fund from the 2020 vintage with between $25 million and $100 million in commitments had 67 LPs, but the the median LP count for funds of this size from the 2025 vintage is 38, the report states.
While the median number of LPs in new funds has declined, the gap between the 25th percentile and 75th percentile for LP count has remained relatively wide across all fund-size intervals, demonstrating some notable variety in the investor makeup of new funds. Among vehicles from the 2025 vintage between $1 million and $10 million in size, a fund at the 75th percentile has 32 LPs; a fund at the 25th percentile has just six LPs.
While the number of LPs in new funds has declined, the median “anchor check,” or largest commitment from a single LP, has held steady. “For funds with between $25 million and $100 million in commitments, the median anchor check has landed at $10 million across each of the past four vintages,” the report notes. “For funds between $1 million and $10 million, the median anchor check has remained consistent at $1 million.”
Faster fundraises
Walker also expects the timeframe between rounds to continue to decline. He cites a wave of AI-driven start-ups that have been able to raise multiple rounds in a single year, which is unusual. “On median, I think the time between rounds is going to come back down a little bit. I don’t know if it’s going to get back down to the old level of 18-24 months, but I think we’re probably past the worst of it. And you see that in the graduation rates [from round to round] a little bit as well.” The percentage of funds raising down rounds is also improving, he adds.
