Venture capital fund benchmarks are the reference points LPs use to judge whether your returns are good: TVPI for total value created, DPI for cash actually returned, and IRR for the speed of those returns. For emerging managers the honest answer is that most published benchmarks don't describe you, because they're built from institutional funds ten to fifty times your size. This guide covers what each metric means, what the reference points actually are, and how to present numbers when your fund is too young to have any.
We see this from an unusual vantage point. VC Lab has accelerated 950+ VC firms across 20 cohorts, representing $7.2B+ in target AUM, and our portfolio has produced 390+ up rounds. That gives us a view of what first-time and second-time funds actually look like, rather than what the top decile of a 2012 vintage looked like.
What TVPI, DPI and IRR actually measure
Three numbers do most of the work in an LP conversation, and they answer three different questions.
TVPI: total value to paid-in
TVPI divides everything your fund is worth, realized and unrealized, by everything LPs have paid in. A TVPI of 2.0x means the fund has created two dollars of value for every dollar called. It's the headline number most managers lead with, and it's the softest of the three, because the unrealized half rests on your own marks.
DPI: distributions to paid-in
DPI counts only cash that has actually left the fund and landed in an LP's account. A fund can carry a 3.0x TVPI and a 0.0x DPI at the same time, and many do. DPI is the number that has become hardest to ignore since 2023, because the exit window narrowed and LPs got tired of paper marks. If you're raising Fund II, expect the DPI question early.
IRR: internal rate of return
IRR annualizes returns and accounts for timing, which makes it the most manipulable of the three. An early markup on a small position can produce a spectacular IRR that means very little. Sophisticated LPs read IRR alongside DPI specifically to catch this.
The three move independently, and that's the point. A fund with high TVPI and low DPI is either young or stuck. A fund with high DPI and modest TVPI probably sold early. Reading any one of them alone tells you almost nothing.
Why standard venture capital fund benchmarks mislead emerging managers
Most benchmark data you'll find comes from institutional sources tracking funds that raised $100MM or more. That's a fundamentally different animal from a $10MM first fund, for reasons that compound.
A large fund needs several very large outcomes to move its multiple. A small fund can hit 3.0x on a single position. That asymmetry means small funds have wider outcome distributions in both directions, so comparing a $10MM fund's interim TVPI to a $250MM fund's is close to meaningless.
Vintage year matters more than fund size for the same reason it always has, and it's rarely disclosed clearly. A 2021 vintage marked in 2023 and a 2021 vintage marked in 2026 are telling you about markdown cycles, not manager skill.
Then there's survivorship. Benchmark datasets are assembled from funds that report, and funds that report are disproportionately funds that are doing well enough to want to. The floor of the distribution is systematically missing.
Our own fundraising data shows how different the small end of the market looks. In our analysis of first VC fund fundraising, the average LP check was $159K and roughly 90% of commitments went to funds under $15MM. That is not a rounding error against institutional benchmarks. It's a different market with different physics.
What the reference points actually look like
Published venture benchmarks tend to cluster around a few widely repeated shapes. Top quartile funds are generally described as clearing 2.5x to 3.0x TVPI over a full life, median funds land closer to 1.5x to 2.0x, and the bottom quartile fails to return capital. IRR reference points for top quartile venture usually sit somewhere in the high teens to mid twenties.
Treat all of those as directional. The sources disagree with each other, the vintages aren't comparable, and the methodologies are rarely disclosed in full. Carta publishes the most useful public cuts through its Data Desk, and AngelList publishes its own platform data. Both are worth reading, and both describe populations that overlap only partially with a first-time $10MM fund.
The more useful frame for an emerging manager is the J-curve. Nearly every fund posts a negative IRR and a TVPI below 1.0x for its first two to four years, because fees are being drawn against a portfolio that hasn't marked up yet. If you're in year two and underwater on paper, you are not failing. You're on schedule.
The metrics beyond the big three
TVPI, DPI and IRR get the attention, but LP diligence questionnaires routinely ask for three more, and not knowing them is an unforced error.
RVPI: residual value to paid-in
The unrealized half of TVPI. RVPI plus DPI equals TVPI, always. Stating RVPI explicitly is a small credibility signal, because it shows you're distinguishing marks from cash rather than hiding behind a blended number.
PIC: paid-in capital
The percentage of committed capital actually called. A fund four years in with 40% PIC is deploying slowly, which is a legitimate question for an LP to ask. Pair it with your investment period end date so the pacing story makes sense.
Gross versus net
Gross returns are before fees and carry. Net is what the LP actually receives. Some managers quote gross without labeling it, and experienced LPs assume the worst when the label is missing. Always specify. The gap between gross and net on a 2 and 20 fund is substantial, and pretending otherwise gets caught.
One more that isn't a metric but functions like one: loss ratio, the share of portfolio companies written off entirely. Venture returns follow a power law, so a healthy fund can carry a high loss ratio and still perform well. Volunteering it signals that you understand your own distribution.
How to present performance when your fund is too young to have any
This is the situation most first-time managers are actually in, and it's where the benchmark conversation usually gets awkward. You can't produce a meaningful DPI in year one. What you can produce is evidence.
Show the track record you do have
Angel investments, SPVs, operator outcomes and board seats all count if you present them honestly. What LPs are testing is whether you can source, pick and access. A clean SPV history with real markups is worth more than a vague claim about a decade in the industry.
Show portfolio construction, not projections
Explain how many checks the fund will write, at what size, at what ownership, with what reserve ratio, and what has to happen for the fund to return 3x. LPs respond to a model they can stress test. They discount a projected IRR to zero.
Show the leading indicators
Before marks exist, the honest signals are portfolio company follow-on rounds, revenue growth in the underlying companies, and co-investor quality. Our own portfolio has generated 390+ up rounds, which is a leading indicator rather than a return, and we describe it that way.
Benchmark your process instead of your returns
When you have no returns to compare, compare everything else: deal flow volume, conversion from first meeting to term sheet, ownership achieved, time to first close. This is where Decile Hub earns its keep, because a fund that can produce clean operating data in month six looks materially more institutional than one reconstructing it from a spreadsheet at year three.
The market context behind today's benchmarks
Benchmarks don't sit still, and the last two years have moved them. Our emerging manager performance data, drawn from a sample of 1,000+ PACTs, 1,000+ LPAs and 900+ funds, found that February through May 2026 ranked among the top five fundraising months in four years, each running 1.2x to 2.2x the same month in 2025.
The composition of who's raising has shifted too. Our generalist versus specialist analysis found generalist funds fell from 22% of new funds in 2020 to 5% in Q1 2026. Solo GPs now account for 61% of new firms in our cohorts, under-40 managers make up 38% versus 25% previously, 28% of GPs are female, and 56% are based outside the United States.
None of that changes what TVPI means. All of it changes which comparison set is fair. A specialist solo GP running a $12MM pre-seed fund out of Southeast Asia is not usefully benchmarked against a generalist $300MM Bay Area Fund IV, and the more the market shifts toward the former, the less the published benchmarks describe the median new manager.
Building your own benchmark set
The practical move is to assemble a comparison set of ten to twenty funds that genuinely resemble yours: same vintage window, same stage, same fund size band, same geography where possible. Track what you can observe publicly, ask peers directly, and be explicit with LPs about what your comparison set is and why.
Saying "against funds of our size and vintage, here's where we sit, and here's the sample" is a stronger position than quoting an industry top quartile figure that everyone in the room knows doesn't apply. It also signals the thing LPs are really evaluating in a first fund, which is judgment.
If you're still assembling the fund itself, fund formation comes first and the benchmark conversation follows. Managers who use the Start Fund to get a first vehicle moving often find that a small realized track record does more for Fund I than any amount of benchmark positioning. And once the fund is live, Decile Partners handles the administration that produces the clean capital account data these calculations depend on.
Frequently asked questions
What is a good TVPI for a first-time venture fund?
There's no reliable published figure specific to first-time funds, and anyone quoting one precisely is probably extrapolating. Over a full fund life, clearing 2.5x TVPI would generally place a fund in strong company. In years one through four, a TVPI below 1.0x is normal and expected because of the J-curve.
Is DPI or TVPI more important to LPs?
It depends on fund age. For a fund under four years old, LPs mostly look at TVPI and portfolio quality because there's nothing else to look at. From roughly year six onward, DPI dominates, because paper marks that never convert to cash have burned a lot of LPs since 2023.
How do I benchmark a fund with no exits yet?
Benchmark the inputs rather than the outputs. Compare deal flow volume, ownership percentages achieved, follow-on rates in your portfolio, and co-investor quality against similar funds. Present these explicitly as leading indicators, not returns.
Why do venture benchmarks vary so much between sources?
Different sources use different fund populations, different vintage groupings, and different treatment of unrealized value, and most don't fully disclose methodology. Survivorship bias affects all of them, because funds that report are funds doing well enough to want to report.
What is the J-curve and how long does it last?
The J-curve describes the shape of fund returns over time. Management fees are drawn from day one against a portfolio that hasn't marked up yet, so the fund shows a negative return early, then recovers as companies mark up and exit. It typically runs two to four years, sometimes longer for funds investing at pre-seed. A fund underwater on paper in year two is on schedule, not in trouble.
Should I show projected IRR in my LP deck?
Generally no. Experienced LPs discount projected IRR heavily and some read it as a signal of inexperience. Show the portfolio construction model and what has to be true for the fund to return capital several times over. Let them do the arithmetic themselves.
Where to go from here
Benchmarks are a tool for context, not a scoreboard you're graded against in year two. Build a comparison set that actually resembles your fund, be explicit about your methodology, and put your energy into the operating data you control. If you're raising a first or second fund and want the structure behind this, VC Lab runs a free 14-week accelerator for new and emerging managers, and our full research library is at the VC Research hub.