Cyrus Reed: Iris, hey — I spent part of the week trying to explain venture capital to my cousin who just got a job offer from a startup, and I realized mid-sentence I couldn't actually explain the core logic. Like why does any rational person fund twenty companies knowing most of them will fail?
Iris Holm: What did you tell her?
Cyrus Reed: Something vague about diversification, which is wrong, and I knew it was wrong even as I was saying it.
Iris Holm: It's not diversification. It's the opposite — it's concentration disguised as a portfolio.
Cyrus Reed: Which is exactly what I want to unpack, because the number that keeps hitting me is — Cambridge Associates found that the top ten percent of companies produced nearly ninety percent of all venture asset-class value. So what is the other ninety percent of companies even doing in these funds?
Iris Holm: Paying for the option to be in the same portfolio as the ten percent. That's it.
Cyrus Reed: Okay but — help me feel what that actually means at the fund level. Because rationally, 'buy things I know will fail' sounds like a terrible strategy.
Iris Holm: Twenty-five lottery tickets. You know twenty-two are worthless before you buy them. But one might pay back more than every ticket combined. Does the math say stop buying tickets?
Cyrus Reed: No — no, it says buy all of them or you might not hold the one that wins.
Iris Holm: Peter Thiel said it clean: in a successful fund, the best investment equals or outperforms the entire rest of the fund combined. Not 'performs well.' Outperforms everything else, summed.
Cyrus Reed: So missing that one company — that's not a bad quarter. That's the fund just... not working. Like structurally not working.
Iris Holm: Structurally not working — and the fee structure is why. VCs take roughly two percent of committed capital as a management fee every year. That's just to operate. The real money, the carry, is twenty percent of profits above a hurdle rate. But carry only matters if the absolute number is large enough to move.
Cyrus Reed: Wait — so what does 'large enough' actually mean? Like in dollars?
Iris Holm: Run the math on a five-hundred-million-dollar fund. A three-times return on a two-million-dollar check — that's four million dollars in profit. Twenty percent carry on that is eight hundred thousand. Sounds fine until you realize the fund needs to return a billion-plus to LPs before carry kicks in at all. That eight hundred thousand is noise.
Cyrus Reed: So a VC is actually better off financially just... ignoring a business that returns three-x?
Iris Holm: Worse than ignoring it — funding it costs them attention and a board seat. A solid, profitable company returning three-x is a distraction. That's the design.
Cyrus Reed: That's a little horrifying? Like — the fee structure is actively selecting against good businesses.
Iris Holm: Bill Gurley put it plainly: venture capital is not a home-run business, it's a grand-slam business. Batting average doesn't matter. Slugging percentage does. You can strike out eighty percent of the time and win if the one swing clears the bases.
Cyrus Reed: Okay but — how do you even know what a grand slam looks like before it happens? Because if you knew, you'd just... fund that one thing.
Iris Holm: You don't know. That's the whole problem. Benchmark Capital looked at Google's Series A in 1999 and passed. Gurley has talked about that miss — Google went on to return something like ten thousand times the investment. One position. Enough to define an entire fund's decade.
Cyrus Reed: Ten thousand — wait, ten thousand x? On a single check?
Iris Holm: Approximately. Which is why you build twenty to thirty positions — not because you believe in each one equally, frankly not even close. You need enough at-bats for the outlier to statistically show up. Miss the Google in your portfolio and the math on everything else is irrelevant.
Cyrus Reed: So they're not picking winners — they're buying enough chances that a winner has to show up somewhere.
Iris Holm: Right. And the asymmetry makes it rational. Downside on any one check is capped — you lose what you put in. Upside is theoretically unbounded. A hundred-x, a thousand-x, ten thousand-x. That asymmetry is the only thing that makes a high-failure-rate portfolio not just acceptable but the correct design.
Cyrus Reed: But then — who decided fund sizes should be five hundred million in the first place? Because it sounds like the grand-slam logic only becomes inescapable once you've already committed to a number that requires it.
Iris Holm: That's exactly where the design choice hides. Fund size is LP expectation hardened into math. But here's what that does to the founder — it cascades all the way down to her desk.
Cyrus Reed: Wait — her desk specifically?
Iris Holm: Picture a founder. Eight months in. Profitable — fifty thousand in MRR, sustainable margins, actual paying customers. Not a rocketship, but real. Her VC shows up for a board meeting.
Cyrus Reed: And that sounds like — okay, that sounds like success? Like a startup that actually works?
Iris Holm: The VC's message: the margins are the problem. Raise another round, hire aggressively, burn cash, capture the market before someone else does.
Cyrus Reed: The margins are — wait, he's saying profitability is bad?
Iris Holm: From the fund's position, yes. A business that exits at fifty million — clean, profitable, founder's proud of it — that check was maybe two million in. Three-x return. We ran that math. It's noise. The fund cannot return LP capital on that outcome. Structurally invisible.
Cyrus Reed: So he doesn't dislike her business — the fund math just can't see it. It literally doesn't register.
Iris Holm: And the pitch he's making — grow at all costs, accept the burn — that's not a personality trait or aggressive management style. It's the only strategy compatible with what the fund needs. Facebook, Google — those are the retrospective proof that makes it feel rational. One player takes the market, everyone else is footnotes.
Cyrus Reed: Right — but Facebook taking social media, or Google taking search, those are winner-take-most markets, aren't they? Like, not every market works that way.
Iris Holm: That's the filter. VCs aren't just looking for big markets — they're looking for markets with network effects or scale economies where the leader takes almost everything. Because only that payoff is asymmetric enough. Andreessen Horowitz built mega-funds specifically to write bigger checks into companies chasing exactly those unicorn-scale outcomes. That's not accidental — it's the power-law logic institutionalized.
Cyrus Reed: And this playbook — it's not new, it's been, what, baked in since when exactly?
Iris Holm: Late nineties. The dot-com era crystallized it — Sand Hill Road firms ran the moonshot playbook, it failed spectacularly for most, but the handful that hit defined the math going forward. The expectation locked in. Which is actually — the question of whether that expectation is mathematical necessity or something VCs just collectively decided to believe — that's a thread we need to pull on, and where it goes is uncomfortable.
Cyrus Reed: Because back to that founder — she knows if she runs the burn strategy and misses the billion-dollar exit, the VC moves to the next fund. She's the one explaining to employees why payroll is short.
Iris Holm: And that's the externalization. The fund doesn't absorb that cost. The founder does. Employees do. The VC is already on to fund four.
Cyrus Reed: Which — okay, but is that a flaw in the model or just... the model working exactly as designed?
Iris Holm: BIP Ventures called it out directly. They said the power law, as practiced by most VCs today, is — their word — BS. Not the math. The practice.
Cyrus Reed: Wait — a VC firm said that? About VC firms?
Iris Holm: Practitioners rarely break rank. That one did. And the distinction they're drawing matters — the mathematical shape of returns is real. The question is whether the behavior that produces it is inevitable or chosen.
Cyrus Reed: So — okay, how do those even come apart? Because if the returns are concentrated, don't you have to behave in a way that chases concentration?
Iris Holm: It gets uncomfortable. In the late 2010s, every major fund piled into on-demand services and fintech simultaneously. Not because each one independently ran the analysis and landed there. Because the power-law logic says back winner-take-most markets — and everyone shared the same logic.
Cyrus Reed: So they're all pattern-matching off the same template, which means — wait, no one is actually doing independent discovery. They're just... collectively creating the thing they're trying to find.
Iris Holm: Valuations inflate. The outlier doesn't materialize at the scale the math required. Sector bubble. That's not the power law predicting outcomes — that's coordinated behavior manufacturing a bust.
Cyrus Reed: And the costs of that bust — I mean, that's founders, employees, anyone downstream. Not the carry.
Iris Holm: Right — and that's the self-fulfilling loop. VCs invoke the power law to justify concentration. Concentration produces winner-take-most dynamics in some markets. Which then gets cited as proof the power law is real. You can't easily separate what the math predicted from what the capital flow caused.
Cyrus Reed: Okay but — the luck question is sitting underneath all of this and I can't get past it. If missing Google's Series A in 1999 is the canonical catastrophe, and Benchmark passed on it — Benchmark, not some random fund — then how much of this is skill? Because if it's mostly luck, the two-and-twenty fee structure is a very confident ask.
Iris Holm: That's the question nobody in the asset class wants to answer cleanly. Canadian empirical data on government-sponsored VCs shows they underperform private VCs on value creation and innovation — but that might just mean governments can't run funds. It doesn't resolve whether private VCs are skilled or just better positioned to absorb luck at scale.
Cyrus Reed: So we genuinely don't know if it's math or mythology — and everyone with carry on the line has a reason not to find out.
Iris Holm: That's the honest place this lands. Whether you think it's math or mythology — the founder still gets the same Tuesday-morning conversation. Growth or death. The mechanism runs whether or not anyone consciously chose it.
Cyrus Reed: And that's — yeah, that's the thing that actually settled for me. Like, I kept wanting to find the moment where someone made the wrong call. But if it's self-reinforcing, if the Cambridge Associates number is real and the carry structure is real, then the founder's situation is the same either way. It doesn't matter if Peter Thiel is describing physics or if a bunch of very smart people just collectively decided this was physics.
Iris Holm: The downstream effect is identical. Middle market starved. Most founders pushed toward strategies that are genuinely incompatible with building a sustainable company. And total VC capital deployed stays almost entirely uncorrelated with how many good startup outcomes you actually get.
Cyrus Reed: Which means — wait, if it's partly chosen, if BIP is right and this is BS as practiced — someone would have to voluntarily write smaller checks, target lower returns, give up the shot at capturing a 50x. And right now there's just... no LP on earth paying two-and-twenty for that.
Iris Holm: No financial reason to choose differently. That's the close. Not the math — the incentive.
Cyrus Reed: Good place to stop. Thanks for walking me through it — I can actually explain it now without lying to my cousin.
