There is an estimated $800 billion sitting in unrealized venture portfolio value across the industry right now. Very little of it has made its way back to the LPs who funded it.
That is the real story of the venture market in 2026. Not the AI boom. Not the secondaries frenzy. The core issue is that a decade of easy money, inflated markups, and delayed exits has created a massive disconnect between what fund managers report and what limited partners actually receive. TVPI is the metric GPs love to show. DPI is the metric LPs are finally demanding.
What DPI and TVPI Actually Mean
TVPI — Total Value to Paid-In — is the headline multiple you see in fund decks. It combines the cash already distributed to LPs with the current marked value of unrealized positions. A fund that called $100M, returned $30M in distributions, and has a portfolio marked at $270M would show a 3.0x TVPI. Looks great. Means almost nothing until those marks convert to exits.
DPI — Distributed to Paid-In — only counts what has actually landed in LP bank accounts. Using the same fund above, the DPI is 0.3x. That fund has returned 30 cents on every dollar called. The rest is paper.
I have sat across the table from GP after GP who leads with TVPI in fundraising decks while burying DPI in an appendix footnote. That gap tells you everything about what they know versus what they want you to believe.
The 2021 Bubble Built a TVPI Mirage
The 2021 vintage is the stress test nobody wanted. Funds raised at peak deployed into companies valued at peak, then marked those positions up again as the market went sideways. Cambridge Associates data shows median VC TVPI for 2021-vintage funds reached 1.4x by end of 2022 — before the real correction hit. By mid-2024, many had drifted back toward 1.0-1.1x while DPI remained nearly flat at 0.05-0.10x.
The 2018-2020 vintage funds are only marginally better. PitchBook's 2024 VC benchmarking data shows median DPI for 2019-vintage funds at just 0.28x — meaning limited partners have received less than 30 cents back on every dollar committed, years into a fund that should be in harvest mode. Top-quartile 2019-vintage funds reached 0.65x DPI. That is considered strong.
Total LP distributions from VC dropped to roughly $60 billion in 2023, the lowest level since 2016, according to PitchBook. The IPO pipeline froze. Strategic M&A slowed. And the marks kept sitting there — on balance sheets, in quarterly reports, in fundraising decks.
Why LPs Are Changing Their Evaluation Framework
LP behavior is shifting measurably. A 2024 Goldman Sachs survey of institutional LPs found that 40% now rank DPI as their primary performance metric — up from 18% in 2020. That is not a marginal shift. That is a structural rewrite of how capital allocators evaluate fund managers.
The reason is practical. University endowments have spending rules — typically 4-5% annual distributions to fund operations. Foundations face similar payout requirements. Pension funds have liability obligations. When VC portfolios lock up capital for 12-15 years while showing only paper gains, these institutions face real cash flow pressure that TVPI does nothing to relieve.
The secondary market has become a pressure valve — and a price discovery mechanism. VC fund stakes were trading at 65-78 cents on the dollar through much of 2023 and 2024. That discount is the secondary market's honest assessment of how much marked NAV will actually convert to distributions. It is not a vote of confidence in the TVPI figures being reported.
The Numbers That Expose the Gap
- •Median DPI for 2019-vintage VC funds is 0.28x as of end-2024, per PitchBook — less than a third of LP capital has been returned
- •Cambridge Associates benchmarks show top-decile funds from the 2015-2017 vintage reaching 2.5-3.5x DPI — the cohort most LPs expected to be the norm
- •VC distributions to LPs totaled approximately $60B in 2023, down from $184B in 2021 — a 67% collapse in actual cash returned
- •40% of institutional LPs now list DPI as their primary evaluation metric for VC managers, up from 18% in 2020 (Goldman Sachs LP survey)
- •VC fund secondaries were bid at 65-78 cents on NAV through most of 2023-2024, implying the market believes roughly 22-35% of marked value won't convert
- •The average time to 1.0x DPI for top-quartile VC funds has extended from 8 years (2010s) to 11-13 years for 2018-2021 vintages
What This Means for Emerging Managers
If you are raising Fund II or Fund III right now, your TVPI from Fund I may be largely irrelevant to sophisticated LPs. What they want to see is whether you have made any distributions at all, whether you have marked anything down honestly, and whether your portfolio has any near-term liquidity catalysts.
The emerging managers who are winning LP conversations in 2025-2026 are the ones who lead with discipline — capped fund sizes, concentrated bets, and genuine exit orientation. They are talking about how they helped portfolio companies get acquired at $50M or $200M instead of riding the "go big or go home" narrative into an indefinite hold.
I have seen this firsthand across my network. LPs who were open to 15-fund relationships in 2021 are now cutting to five or six. The filter is not which funds have the highest TVPI — it is which GPs have proven they can return real capital at a pace that respects LP constraints. That is a DPI conversation, full stop.
The fund that cannot return capital eventually cannot raise capital. TVPI is a story. DPI is proof.
Stay current with VC and startup trends at Value Add VC. Originally published in the Trace Cohen newsletter.