FOUNDATION SERIES FOR FAMILY OFFICES · No. 1 of 5
How a Venture Fund Actually Works
The vehicle most professional venture investing runs through — explained plainly, with no product attached.
Pascal Bouvier, MiddleGame Ventures · June 2026
We have been asked how venture funds actually work, many times by family office executives, or the family itself as they had not yet invested in such an asset class. This article is my attempt at demystifying the mechanics of a venture fund.
Most family offices meet venture capital sideways. A friend, a fellow principal, or a banker brings a single company. Before any single company is worth an opinion, the vehicle that most professional venture investing runs through — the fund — deserves to be understood on its own terms. The mechanics are not complicated, but they are rarely set out for a principal without a term sheet in the way.
A partnership with a clock
A venture fund is a partnership with two sides and a fixed life. On one side sit the limited partners — you, the family office — who commit the capital. On the other sits the general partner, the firm, which invests and manages it. The feature that surprises newcomers most is that you do not hand over the money on day one. You make a commitment: a promise to fund up to an agreed amount, which the GP then draws down over years. And the whole arrangement is time-boxed — a venture fund typically runs about ten years, often with a year or two of permitted extensions.[1] That clock governs everything that follows.
How the fund comes together: closings and the legal spine
A fund does not open with all its capital in the room. It holds a first close once enough commitments are secured to begin investing, and additional LPs join at rolling closes thereafter, up to a final close that conventionally falls no more than twelve months after the first. LPs who come in at a later close usually pay an equalization true-up — their share of costs since the first close, plus interest — so that the earliest investors are not disadvantaged for having moved first.
Three documents govern the relationship. The Limited Partnership Agreement (LPA) is the fund’s constitution: it sets the economics, the powers of the GP, the investment remit, the governance and the life of the fund. Each LP accedes to it by signing a deed of adherence — the subscription document by which you formally join the partnership on the LPA’s terms. And a large or strategically important LP may negotiate a side letter: a bilateral agreement granting specific rights — bespoke reporting, fee terms, most-favoured-nation protection, excuse rights, and the like — that sit on top of the LPA without rewriting it for everyone.
One right worth understanding because it can be valuable to a family office: co-investment. A firm will sometimes grant co-investment rights — the option to invest directly alongside the fund in a particular company, usually with little or no additional fee or carry — typically to an LP that brings genuine strategic value. It is a meaningful sweetener, and a subject we treat in its own right later in this series. Governance sits alongside the economics: most funds run an LP Advisory Committee of larger investors to rule on conflicts and valuation questions, and a key-person clause that pauses new investment if the named partners stop devoting their time to the fund.
Money in: what you actually pay, and when
When you commit, say, €5 million, very little leaves your account at first. Money is drawn in two distinct streams,. The first stream is the cost of running the fund — the management fee, conventionally around 2% a year, charged on committed capital during the investment period and typically stepping down afterward,[2] together with fund-level expenses such as administration, audit, custody and legal. Both of these are called periodically, usually once a year. The second stream is investment capital: the money that actually buys companies, which is called deal by deal, as and when an investment is made. Your commitment, in other words, is put to work gradually, not parked.
This split has a consequence every principal should internalize, : the capital actually available to invest in companies is your total commitments minus the lifetime cost of management fees and fund-level expenses. A €100 million fund does not put €100 million into startups — closer to €80–85 million reaches companies, the rest having paid to run the vehicle over a decade. the principle holds: net invested capital is always less than committed capital,.
What the GP earns: carry, the hurdle, and the waterfall
The management fee keeps the lights on;. The reward is carried interest, conventionally 20% of the fund’s profits — but earned only after the LPs come first. When capital returns, it flows in a set order, the distribution waterfall: first your invested capital is returned to you; then your preferred return, or hurdle, most often around 8% and in some funds as high as 12%;[3] then the GP “catches up”; and only thereafter do profits split, conventionally 80% to LPs and 20% to the GP. In Europe this is almost always computed on a whole-fund basis — the GP earns carry once the entire fund is in profit, not deal by deal. The sequence is the alignment: the GP is paid to return your money first and to share in the surplus second.
Alignment runs the other way too. The GP invests its own capital in the fund — typically at least 1–2% of its size[4] — so the people managing your money lose alongside you when they are wrong.
Two halves of a fund’s life: new bets, then reserves
A fund’s life divides cleanly in two. During the investment period — the first three to five years — the GP makes its new investments, building the portfolio of names it has chosen to back. After it closes, in the post-investment period, no new companies are added; capital is deployed instead as follow-ons into the existing portfolio — defending or increasing ownership as the best companies raise their next rounds.
This is why reserves matter more than newcomers expect. A disciplined fund deliberately holds back a large share of its capital — often 30–35% of the total, sometimes more — precisely to fund those follow-ons. Reserves are not idle caution; they are the mechanism by which a manager doubles down on the few companies that bend the curve (more on that curve below).
The J-curve: why the early years look like a loss
For its first few years a fund usually shows a paper loss, and this is by design, not distress. Fees and expenses are charged from the start; companies are bought at cost and held there until some event re-prices them; and the winners take years to mature. Plotted over time, value dips before it climbs — the shape allocators call the J-curve. A principal who reads the year-three dip as failure has misread the instrument. The curve has simply not yet been given time to do its work.
How returns are actually made: the power law
This is the single most important idea in the piece, because it is the opposite of how most family balance sheets are built. Venture returns do not come from a basket of companies each doing modestly well. They come from a very small number of extreme winners that pay for everything else. In one of the most-cited data sets — more than 21,000 U.S. financings between 2004 and 2013 — about 65% of venture investments returned less than the capital put in, only around 4% returned more than ten times, and a sliver, roughly the top 0.4%, returned more than fifty times.[5] That distribution is the power law, and it forces two truths a family office must hold at once. Loss is normal: being wrong often is not,— it is the cost of being positioned for the rare outlier. And the entire result depends on access to those outliers, and on having the reserves to keep backing them. A fund’s job is not to avoid losers; it is to own enough of the handful of companies that bend the curve.
How you measure it: DPI, TVPI, IRR
Three numbers describe a fund’s performance,
DPI — distributed to paid-in. Cash actually returned to you divided by cash you put in. It is the only figure that has already happened. A DPI of 1.0x means you have your money back; anything above that is real profit in hand.
- TVPI — total value to paid-in. DPI plus the GP’s estimate of what the still-held companies are worth. Useful as direction, but part opinion — paper marks, not cash.
- IRR — internal rate of return. The annualized, time-weighted return. Sensitive to timing and easily flattered in a fund’s early years.
The discipline is simple. Early in a fund’s life, treat TVPI and IRR as direction, not verdict; over the full life, DPI is the number that pays for things. Marks are a story; distributions are the fact. Because TVPI leans entirely on those marks, ask how a GP values its unrealized positions — credible managers follow recognized standards such as the IPEV valuation guidelines — and ask for realized DPI by vintage before discussing anything else.
The risks, named plainly
Venture capital carries four risks a family office must accept with open eyes. It is illiquid: your capital is committed for roughly a decade, with no redemption window. It is a blind pool: you commit before the companies exist, so you are underwriting the manager’s judgment, not a fixed list of assets. It is loss-heavy at the level of individual deals, as the power law makes plain. And it carries a standing obligation: when the GP calls capital you must fund it, and the LPA’s default provisions for failing to do so can be severe, up to forfeiting part of your interest..
Why this matters for a family office
None of these features makes venture a poor fit for a family office. They make it a specific instrument with a specific shape: patient, illiquid, manager-dependent, and driven by outliers. The family office — with permanent capital and a horizon measured in generations — is unusually well suited to that shape, in some respects better suited than the institutions that dominate the asset class. Whether to gain the exposure through funds, through direct deals, or through both is the subject of the next piece. The mechanics above are simply the price of entry to that conversation.
Next in the series: “Funds, Direct, or Both — How Family Offices Really Get Paid in Venture.”
[1]A ten-year fund term, with provision for one or two one-year extensions, is the long-standing market convention for venture funds. See VC Lab, “VC Fund Mechanics,” govclab.com.
[2]A management fee of roughly 2% a year is the industry norm — typically charged on committed capital during the investment period and stepping down thereafter (often to invested capital) — alongside 20% carried interest: the “2 and 20” model. See Qubit Capital, “Two and Twenty.”Any specific fund’s economics are set out in its own documents.
[3]A preferred return, or hurdle, must be returned to LPs before the GP earns carry. It is most commonly set around 8%, and in some funds as high as 12%; ranges vary. See Carta, “Hurdle Rate.”
[4]A GP commitment of at least 1–2% of fund size — the firm investing its own capital alongside its LPs — is the market norm and a basic test of alignment. See VC Lab, “VC Fund Mechanics.”
[5]Correlation Ventures’ analysis of more than 21,000 U.S. venture financings (2004–2013), as widely reported: about 65% returned less than the capital invested, roughly 4% returned more than 10x, and the top ~0.4% returned more than 50x. See, e.g., The VC Factory, “The Super Power Law.”
