Data sourced from Dakota Private Markets, the private fund performance platform powered by Dakota. Learn More | Request Access
The median private equity fund carries a TVPI of 1.53x. The median 2021-vintage private equity fund has a DPI of 0.16x. Both are correct, and together they say something neither says alone: roughly 90% of that fund's reported value has not been realized yet.
That gap is RVPI, the multiple almost nobody publishes a benchmark for. In this article, we'll cover what the three multiples measure, what the averages look like by strategy and vintage, and how to calculate the residual the tables do not show.
The Three Multiples in One Line Each
All three share the same denominator, paid-in capital, which is what makes them additive.
|
Multiple
|
Formula
|
Measures
|
|
TVPI
|
(Distributions + Remaining Value) ÷ Paid-In
|
Total value created per dollar called
|
|
DPI
|
Distributions ÷ Paid-In
|
Cash actually returned
|
|
RVPI
|
Remaining Value (NAV) ÷ Paid-In
|
Value still held in the portfolio
|
The identity that governs all three:
TVPI = DPI + RVPI
This always holds. A fund reporting 1.53x TVPI and 0.58x DPI is carrying 0.95x in unrealized NAV, whether or not it prints that number on the page. For the full definitional treatment of these and the metrics around them, see the private fund performance data dictionary.
Average TVPI by Strategy
Median TVPI across 1,757 funds tracked in Dakota Private Markets, Q1 to Q2 2026. Note the quartile convention: first quartile is the top performers, third quartile the bottom.
|
Strategy
|
Median TVPI
|
1st Quartile
|
3rd Quartile
|
Quartile Spread
|
Funds
|
|
Private Equity
|
1.53x
|
1.91x
|
1.23x
|
0.68x
|
1,020
|
|
Venture Capital
|
1.41x
|
1.89x
|
1.16x
|
0.73x
|
159
|
|
Real Assets
|
1.33x
|
1.60x
|
1.09x
|
0.51x
|
193
|
|
Private Credit
|
1.28x
|
1.47x
|
1.12x
|
0.35x
|
385
|
The medians are close together. The spreads are not, and that is the more useful column.
Venture capital shows the widest gap between top and bottom quartile at 0.73x, private credit the narrowest at 0.35x. A private credit manager in the third quartile at 1.12x is much closer to the median than a venture manager in the same position. Manager selection carries roughly twice the consequence in venture that it does in credit, which is the argument for concentrating diligence effort where the dispersion actually is.
One caution on reading the maximums. Private credit's high end reaches 9.33x, well past private equity's 5.57x, which reflects a small number of distressed and special situations funds rather than anything typical of the asset class.
Median DPI by Asset Class and Vintage
DPI only means something against fund age, so this one has to be read by vintage rather than as a single number.
|
Vintage
|
Private Equity
|
Venture Capital
|
Private Credit
|
Private Real Estate
|
Real Assets & Infra
|
|
2023
|
0.07x
|
0.07x
|
0.10x
|
0.07x
|
0.10x
|
|
2022
|
0.09x
|
0.17x
|
0.28x
|
0.10x
|
0.11x
|
|
2021
|
0.16x
|
0.28x
|
0.34x
|
0.22x
|
0.12x
|
|
2020
|
0.30x
|
0.20x
|
0.47x
|
0.12x
|
0.16x
|
|
2019
|
0.48x
|
0.32x
|
0.59x
|
0.24x
|
0.49x
|
|
2018
|
0.58x
|
0.15x
|
0.88x
|
0.57x
|
0.55x
|
|
2017
|
0.79x
|
0.65x
|
0.95x
|
0.58x
|
0.86x
|
|
2016
|
1.21x
|
0.93x
|
1.02x
|
1.06x
|
0.57x
|
|
2015
|
1.23x
|
1.20x
|
1.07x
|
0.93x
|
1.06x
|
Three things stand out.
- Private credit returns capital first, and the lead widens in recent vintages. At the 2018 vintage, median credit DPI runs 1.5x private equity's. By 2021 it runs 2.1x. Contractual cash flow does not wait for an exit window, which is most of the explanation.
- 2016 is the crossover. Private equity, private credit, and private real estate all cross 1.0x median DPI at the 2016 vintage, meaning a decade is roughly what it has taken for the median fund to return the capital it called.
- Venture is not a curve. The 2018 vintage sits at 0.15x, below both 2019 and 2021. Venture realizations depend on a small number of exits landing, so cohort medians move in steps rather than a smooth progression. Read venture DPI by individual vintage or not at all.
These are medians. Your fund is not a median. A benchmark is only actionable when the peer group matches on vintage, strategy, geography, and fund size. Dakota Private Markets holds the fund-level records behind every figure above, request access.
RVPI: The Multiple Nobody Benchmarks
There is no published median RVPI table, here or anywhere, and that absence is itself informative. RVPI is the least comparable of the three because it depends entirely on manager valuation marks, and marks are struck on different policies by different managers at different times.
You can still derive it, because the identity holds at the fund level:
RVPI = TVPI − DPI
Applied to a private equity fund sitting at the strategy median TVPI of 1.53x:
|
Vintage
|
Median DPI
|
Implied Residual
|
Share of Value Unrealized
|
|
2015
|
1.23x
|
0.30x
|
20%
|
|
2018
|
0.58x
|
0.95x
|
62%
|
|
2021
|
0.16x
|
1.37x
|
90%
|
|
2023
|
0.07x
|
1.46x
|
95%
|
Read this as illustration, not benchmark. It applies one all-vintage median TVPI to each vintage's DPI, and a 2023 fund does not actually carry 1.53x TVPI. Medians are also not additive, so subtracting one median from another does not produce a true median RVPI. What the table shows correctly is the shape of the problem: for anything raised after 2020, the overwhelming majority of reported value is still a mark rather than a distribution.
The practical consequence: when a 2021-vintage fund presents a strong TVPI, close to all of it is unrealized, and the diligence question is not what the multiple is but who struck the valuation and on what policy. That is also why lack of standardized reporting makes private fund benchmarks so difficult.
How to Read the Three Together
- Lead with DPI, matched to vintage. It is the only one of the three a manager cannot influence through valuation policy.
- Read the TVPI quartile spread before the median. The median tells you the middle. The spread tells you what manager selection is worth in that strategy.
- Treat RVPI as exposure, not upside. It is the portion of the story that depends on future exits clearing current marks.
A 2023 fund at 0.07x DPI and a 2016 fund at 1.21x are both performing normally. Put them in one table without the vintage column and one of them looks broken.
Medians are a starting point, not an answer. Dakota Private Markets holds the fund-level records behind every figure above: more than 18,000 funds and nearly 160,000 performance records across seven asset classes, with Net IRR, TVPI, DPI, and RVPI captured on every record.
Build the peer group your comparison actually requires by filtering on vintage year, asset class, sub-strategy, geography, and fund size, then add portfolio company sector, a cut most benchmarking tools cannot make. Coverage runs deepest where the capital is: roughly 5,300 private equity funds, 4,600 private real estate, 1,900 private credit, 1,300 real assets, 1,150 venture capital, and 600 private infrastructure, plus evergreen and interval funds.
Fund performance is one of four datasets under a single login. The others are GP and sponsor intelligence covering investment theses, portfolio companies, deal activity, and team detail; private company transaction data spanning sponsored and non-sponsored businesses with entry multiples and deal structures; and custom benchmarking that lets you construct a peer group rather than accept a category average.
Every record is reviewed by Dakota's research team before publication rather than auto-populated from filings, the data updates daily, and filtered results export to Excel or CSV for the diligence file or the IC memo.
Request access to Dakota Private Markets.
