Until now, the venture capitalist has been the person across the table from the founder. This module goes inside the firm. A VC firm is not a company in the normal sense — it's a fund (or series of funds) with a peculiar structure, a fixed lifespan, an unusual fee model, and a set of investors of its own who almost never appear in the startup story. Understanding how the firm actually works — how it's structured, how it makes money, who its money comes from, and how its partners are made and unmade — explains a great deal of venture-capital behavior that otherwise looks irrational.
The first thing to understand about a venture-capital firm is that "the firm" and "the fund" are different things, and the distinction explains almost everything else. When you read that "Sequoia invested in Stripe," the entity doing the investing was not the firm called Sequoia Capital — it was a specific fund (say, Sequoia Capital XII, L.P.) that Sequoia's partners manage. The firm is the durable brand and team; the funds are the specific pools of capital, each with its own investors, lifespan, and accounting.
A venture fund is almost always structured as a limited partnership — a legal form with two kinds of participants:
The GP is the venture firm — the partners who source deals, make investment decisions, sit on boards, and manage the portfolio. The GP makes all investment decisions and runs the fund's day-to-day operations. In exchange, the GP earns a management fee and a share of the profits (carried interest). The GP also typically commits some of its own capital to the fund (the "GP commit"), which aligns its interests with the investors.
The LPs are the investors who provide nearly all the money. They are pension funds, university endowments, sovereign wealth funds, insurance companies, family offices, and funds-of-funds. LPs are "limited" in two senses: their liability is limited to the capital they commit (they can't lose more than they put in), and they are limited in control (they cannot direct the fund's investment decisions — that's the GP's job). LPs are passive capital providers who trust the GP to invest well.
Money flows in two layers. Layer one: LPs commit capital to a fund managed by a GP. Layer two: the GP deploys that capital into startups. The founders you've been reading about for eight modules interact only with layer two. Layer one — the LPs and the fund structure — is invisible to founders but drives nearly all GP behavior: the fund's size, its lifespan, the pressure to deploy and to return capital, and the urgency around exits all originate in the LP relationship.
This structure is why venture capitalists behave the way they do. A VC pushing a portfolio company toward a faster exit, or pressing for a bigger outcome, or declining to fund a perfectly good business that can't become huge — these behaviors trace back to obligations the GP has to its LPs in layer one, obligations the founder never sees directly.
A venture fund is not a permanent institution. It is a temporary vehicle with a defined lifespan — almost always around 10 years, sometimes with provisions to extend by a year or two. This finite lifespan is the clock that governs everything the GP does, and understanding it explains a great deal of venture timing behavior.
A typical fund life splits into two periods:
During the first 3-5 years, the GP actively makes new investments, deploying the committed capital into a portfolio of startups. New companies are added to the portfolio during this window only. Once the investment period ends, the fund makes no new initial investments — it only follows on into existing portfolio companies.
During the back half, the GP works the existing portfolio — supporting companies, making follow-on investments, and pursuing exits. The goal of this period is to turn the portfolio into cash (via IPOs, acquisitions, or secondary sales) and return it to the LPs. By year 10, the fund is ideally fully liquidated.
The 10-year clock creates structural pressures that show up everywhere in venture behavior:
Venture firms earn money in two distinct ways, traditionally summarized as "2 and 20": a 2% annual management fee on committed capital, and 20% carried interest on the fund's profits. The two are economically very different, and the distinction shapes GP incentives in important ways.
The management fee covers the firm's operating costs: partner and staff salaries, office rent, travel, legal, the platform team (from Module 04). On a $200M fund, a 2% fee is $4M per year — enough to run a small partnership. Crucially, the management fee is paid regardless of whether the fund makes money. It's the GP's salary, guaranteed. Over a 10-year fund life, 2% per year sums to roughly 20% of the fund consumed by fees (often structured to step down in later years).
Carried interest is the GP's share of the fund's profits — typically 20% of everything the fund returns above the LPs' original capital. If a $200M fund returns $600M, the profit is $400M, and the GP's 20% carry is $80M. Carry is where venture partners actually build wealth; the management fee just keeps the lights on. Carry is paid only if the fund profits, which (in theory) aligns the GP with the LPs — both make real money only if the investments succeed.
Many funds include a hurdle rate (also called a preferred return) — LPs must receive their capital back plus a minimum return (often 8%) before the GP earns any carry. Some funds have a catch-up provision that lets the GP "catch up" to its full 20% once the hurdle is cleared. And nearly all funds return capital through a distribution waterfall: LPs get their capital back first, then the preferred return, then the GP catches up, then remaining profits split 80/20. The structure ensures LPs are made whole before the GP collects its profit share.
"2 and 20" is the traditional baseline, but the actual numbers vary. Top-tier firms with strong track records sometimes command 25% or even 30% carry (Sequoia, Benchmark, a16z have charged premium carry at various points). Some large multi-stage firms charge lower management fees on enormous funds (because 2% of a $5B fund is $100M/year, far more than needed to operate). Newer or smaller firms sometimes offer LPs better terms to attract capital. The "2 and 20" shorthand is a useful baseline but the real economics are negotiated fund by fund.
The capital that funds the entire venture ecosystem ultimately comes from the limited partners. Founders rarely think about LPs, but every venture check traces back to one. Understanding who they are and what they want explains the constraints GPs operate under.
| LP type | What they are | What they want |
|---|---|---|
| Pension funds | Public and corporate retirement funds (CalPERS, CPPIB) | Long-term returns to meet pension obligations; venture is a small high-return slice |
| Endowments | University and foundation endowments (Yale, Harvard, MIT) | Long-horizon growth to fund institutional spending in perpetuity |
| Sovereign wealth funds | State investment vehicles (PIF, GIC, Temasek, Mubadala) | Returns plus strategic/geopolitical objectives |
| Insurance companies | Life and general insurers with long-dated liabilities | Returns to match long-dated obligations; conservative |
| Funds-of-funds | Funds that invest in other funds (allocators) | Diversified access to venture for smaller institutions |
| Family offices | Private wealth-management vehicles for wealthy families | Returns plus often direct co-investment opportunities |
What unites these LPs: they have long time horizons (they can lock up capital for a decade), large balance sheets (a venture allocation is a small slice of a huge portfolio), and a need for the high returns that top-quartile venture can provide. What they want from a GP: top-quartile returns, honest reporting, and access to the GP's future funds. The relationship is long-term — an LP that commits to Fund III usually expects to be offered Fund IV.
The single most influential development in venture LP history was the "endowment model" pioneered by David Swensen at Yale in the 1990s. Swensen argued that long-horizon institutions like university endowments were uniquely suited to illiquid, high-return assets like venture capital — they didn't need liquidity, so they could capture the "illiquidity premium" that venture offered. Yale's allocation to venture and private equity, far higher than conventional wisdom allowed at the time, produced decades of market-beating returns and was copied by endowments worldwide. The endowment model is a major reason venture capital grew from a cottage industry into a meaningful institutional asset class.
One structural feature does a lot of work to align the GP with its LPs: the GP commit. The general partner is expected to invest a meaningful amount of its own money into its own fund — traditionally about 1% of the fund's size, though increasingly 2-5% for funds that want to signal strong alignment.
The logic is simple and powerful: if the partners have their own wealth at risk in the fund alongside the LPs, they will (in theory) make better decisions than if they were investing only other people's money. The GP commit converts the relationship from "we manage your money for a fee" into "we're in this alongside you." A GP that commits 5% of a $200M fund has $10M of its own capital at risk — enough to feel real losses.
The management fee, paid regardless of performance, creates a baseline misalignment — the GP gets paid even if every investment fails. Two mechanisms pull the other way: carried interest (the GP earns its real money only if the fund profits) and the GP commit (the GP loses its own money if the fund fails). LPs scrutinize both. A GP with a large personal commit and a track record of letting carry, not fees, drive its economics is structurally better aligned with its investors than a fee-maximizing operation.
The GP commit also matters for what it signals. An LP evaluating a new fund pays close attention to how much of the partners' own money is going in. A small commit ("we'll put in the minimum 1%") signals lower conviction than a large one ("the partners are putting in 5% of the fund from their personal wealth"). For newer GPs without a track record, a substantial personal commit is one of the few credible signals of conviction available.
A venture firm doesn't raise one fund and stop. It raises a series of funds over its lifetime — Fund I, Fund II, Fund III, and so on — each a separate 10-year vehicle with its own LPs (often overlapping) and its own portfolio. At any given moment, a successful firm is typically managing several funds simultaneously, each at a different point in its lifecycle. This creates what's often called the fundraising treadmill.
Consider a firm in steady state. It might be:
The treadmill exists because of the fund clock. A firm that waits until Fund III is fully deployed before raising Fund IV would have a gap with no capital to invest — and would lose deal momentum, partner continuity, and LP attention. So firms raise the next fund while the current one still has dry powder, typically launching the new fund when the prior one is 60-80% deployed. This means the partners are essentially always either deploying or fundraising, often both.
The treadmill explains several otherwise-puzzling GP behaviors:
A venture firm's internal structure is its own world, largely invisible to founders. Understanding the career structure helps explain who you're actually dealing with when you meet "a partner at the firm," and why the individual matters as much as the firm (a point from Module 04).
A typical venture firm has several tiers, though titles vary:
Carry — the real wealth in venture — is allocated among the firm's partners, and the allocation is a perpetual source of internal politics. Senior partners take the largest carry shares; junior partners earn into larger shares over time. Carry typically vests over the fund's life (often 4-5 years), which means a partner who leaves mid-fund forfeits unvested carry. This vesting is the golden handcuff of venture — a partner with large unvested carry in a strong fund has a powerful financial reason to stay.
The making and unmaking of partners is ultimately about deal attribution. A partner who led the investment in a fund-returner (Module 04's power-law math) becomes powerful and wealthy; a partner whose deals all failed quietly leaves. Venture is unusually meritocratic in this narrow sense — the returns are measurable, attribution is (somewhat) traceable, and a partner is ultimately only as valuable as their best investments. This is also why partners fight over deal attribution: in a business where one fund-returner can define a career, getting credit for the right deals is existential.
To make the fund economics concrete, trace $200M of committed capital through a complete fund lifecycle. This shows where the money actually goes — to fees, to investments, and ultimately back to LPs and GPs.
A firm raises a $200M fund (Fund III). Terms: 2% annual management fee for the 10-year life, 20% carried interest, an 8% preferred return (hurdle) to LPs, and a GP commit of 2% ($4M of the partners' own money, counted within the $200M).
The first thing to notice: not all committed capital gets invested. Roughly 16-20% is consumed by fees over the fund's life. The GP deploys what remains — about $168M — into portfolio companies.
Using the power-law distribution from Module 04, the $168M deployed across ~25 companies produces a gross return. Suppose the portfolio returns 3.5× gross on invested capital — a good-but-not-spectacular outcome:
Now the $588M flows back through the waterfall. LPs get their capital back first, then their preferred return, then the GP catches up, then the remainder splits 80/20.
| Waterfall step | Amount | Goes to |
|---|---|---|
| 1. Return of capital | $200M | LPs (their committed capital back) |
| 2. Preferred return (simplified) | ~$80M | LPs (the 8% hurdle, roughly) |
| 3. Remaining profit | $308M | Split 80/20 (after GP catch-up) |
| 4. GP carried interest (20%) | ~$78M | GP (carry on profits) |
| 5. LPs total | ~$510M | LPs (capital + preferred + 80% of remaining) |
Several things worth absorbing from the numbers:
A venture firm is a machine for converting LP capital into startup investments and (ideally) returning multiples of that capital, taking fees along the way and a large slice of the profits at the end. The 10-year clock, the 2/20 economics, the GP commit, the fundraising treadmill, and the partner career structure all exist to make this machine work — and to align (imperfectly) the GP's incentives with the LPs who provide the money. Every venture behavior a founder finds puzzling traces back to some feature of this machine.
Six questions on fund structure, the 2/20 economics, the LP relationship, and the institutional pressures that shape GP behavior.