Venture capital is an instrument for buying variance, not for rewarding excellence. A fund returns money to its investors when a tiny handful of its bets pay off enormously, while the median position is expected to lose. That one fact, not greed or short-termism or a bad partner across the table, is why a genuinely excellent business can be the worst possible thing to finance with venture money. If your company would grow 25% a year for a decade and throw off cash the whole time, it is, to the fund, a failure that happens to be pleasant to look at.
Feel the arithmetic, because the pressure a VC puts on you is not a personality trait. It is math wearing a suit.
The number every check has to clear
Take a $100M fund. Over its roughly ten-year life, about 20% is consumed by management fees at 2% a year, so on the order of $80M actually gets invested across perhaps 25 to 30 companies, with a chunk held in reserve to follow on into the ones that work. For that fund to be considered good, it needs to return something like three times the committed capital, so it has to produce on the order of $300M of realized value.
That $300M does not arrive evenly. Across thirty investments, most return under 1x, several go to zero, and one or two return the majority of the entire fund. This is the empirical shape of venture returns, the power law that Chris Dixon popularized from the Horsley Bridge data: a small fraction of investments produce the overwhelming bulk of the gains, and the best funds are not the ones that avoid losers but the ones that catch a bigger winner. So every partner underwrites each check as if it individually needs a credible path to returning the whole fund, because statistically only one or two of them will.
Now add the ownership constraint. A fund typically owns 10 to 20% of a company at exit after dilution. To get $300M back from a single winner at 15% ownership, that company has to exit for roughly $2B. So the real question a partner is asking, whether or not they say it out loud, is whether this company has a believable path to a billion-plus outcome. If it doesn't, it cannot be the investment that carries the fund. At best it can be one of the many that don't matter.
Your 25%-a-year cash machine, with a realistic ceiling around $200M of enterprise value, returns the fund about $30M at 15% ownership. Against a fund that needs $300M, that is a rounding contribution. It also consumed a board seat, a partner's attention, and reserve capital that were all supposed to be aimed at a fund-returner. To you, $30M is generational wealth. To the fund it is an opportunity-cost failure that also refuses to die and free up the slot.
Why they push, and why the push can kill you
Given that math, watch what becomes rational for your investor. Consider two versions of your company:
- Version A: 90% chance of being worth $150M, 10% chance of zero.
- Version B: 30% chance of being worth $2B, 70% chance of zero.
The fund strictly prefers B. Version A's expected value is $135M and, more damningly, can never return the fund at any ownership stake the fund realistically holds. Version B's expected value is $600M and clears the bar. A diversified investor holding thirty of these prefers the coin flip every single time, because across the portfolio the 70% zeros wash out and the rare winners pay for everything, including the losers.
You are not diversified. You hold exactly one position: your company, your decade, your reputation, often your savings. On Version B you personally face a 70% chance of zero. The identical gamble that is obviously correct for the fund is close to reckless for you. This is the principal-agent problem in its cleanest form. Your investor is not a bad agent. You and your investor simply have different utility functions over the same outcome distribution, because one of you is holding thirty tickets and the other is holding one.
Bet-sizing theory makes the divergence precise. The fraction of your bankroll you should stake on a positive-expectation gamble scales with how many independent bets you get to make. The fund, making thirty, can size each one aggressively. You, making one, should size it far more conservatively, and "the company itself" is a bet you cannot fractionally size once you have raised on the premise that you will swing for B.
The mandate to swing is not delivered as a speech. It is encoded in the paper. Preferred stock with a liquidation preference gets the fund its money back before common holders see a dollar, and a senior, participating stack means a modest exit can make the fund whole while leaving the founder with little. The instrument is deliberately tuned so the fund is nearly indifferent to the small good outcome and strongly prefers the swing. Anti-dilution ratchets and pro-rata rights point the same direction. When you sign a term sheet, you are not just taking money. You are adopting the fund's utility function as your company's governing law.
So the VC pushes, correctly by their lights: raise more, spend into the tail, widen the TAM narrative, treat capital as a weapon. Every one of those moves raises the value of the winning branch and the probability of the zero branch at the same time. For a portfolio, that is right. For your single, undiversified life, it can be ruinous. Nobody has to be acting in bad faith for it to end with your good business dead because it was governed into chasing an outcome its structure could never produce.
The diagnostic: fund-returner, or great business with a ceiling
Two questions, in order.
First: is your credible best case genuinely a fund-returner? Not your pitch-deck best case. The one you would bet your own money on. That requires winner-take-most dynamics, a market where getting large first makes you durably hard to unseat, so capital buys a moat instead of merely renting a lead. This is where most founders deceive themselves, because most claimed moats do not survive contact. I've argued that network effects are the most overclaimed moat in the business: they are usually local, saturating, and multi-tenant, which means being first and biggest buys far less durability than the winner-take-all story needs. If that describes your moat, your realistic ceiling is bounded no matter how much you raise. Capital does not convert a ceilinged business into a fund-returner. It converts it into a ceilinged business now carrying a 70% chance of zero.
Second, and separately: even if you could be a fund-returner, do you want what the money brings? Venture capital is not neutral fuel. It arrives with a clock, because a fund has to return capital in roughly seven to ten years and you are now on someone else's liquidity schedule. It arrives with a mandate to grow into the tail or die trying, because the profitable middle is worthless to the fund. And it arrives with a transfer of control: the board that will, if you stall, replace you or force a sale at the moment that is right for the fund's timeline rather than your life. Founders routinely answer the first question yes and never ask the second. Then they are surprised when a profitable, growing company gets pushed to raise again and torch its own profitability, because "a good outcome" and "a fund-returning outcome" were never the same target.
The math is moving under everyone's feet
The fund-returner calculation is not static, and right now it is being repriced in real time across AI. A large share of AI-application companies raised at valuations underwritten on software economics: near-zero marginal cost, gross margins north of 80%, and therefore the kind of terminal value that clears the billion-dollar bar. Much of their cost of goods, though, is inference, and inference is a metered input with a falling and volatile price. I've made the case that the collapse in inference cost is going to break the pricing model of most AI companies: when your input cost drops an order of magnitude, competitors reprice, customers expect the savings passed through, and the fat margin that justified the valuation compresses toward a commodity passthrough.
When that happens, terminal value shrinks, and a company underwritten as a fund-returner quietly stops being one, mid-flight, after it has already taken the capital and the mandate and the clock. The founder is now legally and structurally committed to swinging for a fence that receded. That is the worst position in this entire essay: bound to Version B governance in a market that just became Version A.
The honest counterargument
The steel-man is real. When a market is genuinely winner-take-most, being undercapitalized is the ruinous choice, and the VC's push is aligned with your survival rather than opposed to it. If there truly is no stable, profitable, medium-sized equilibrium, if the market resolves to one or two winners and everyone else dies regardless, then raising hard and spending into the tail is not recklessness. It is the only strategy with a non-zero survival branch. Losing slowly and solvently is still losing. In those markets, the founder who refuses venture money out of a preference for control is choosing a dignified zero. Blitzscaling exists because some markets actually reward it.
So one question decides everything: does a stable middle exist? Is there a version of your company that is medium-sized, profitable, and durable, one that can sit at $30M of revenue, growing 20%, for a decade without being killed by a better-capitalized entrant? If that version exists, venture money's pressure is destroying optionality you actually hold, and you should be extremely reluctant to sell it. If that version does not exist, if the physics of your market genuinely forbid a stable middle, then the VC is not imposing the coin flip on you. The market already did. The capital is just the instrument for playing the hand you were dealt.
I have had to run this test honestly on my own companies. Cash-on-delivery commerce in trust-scarce markets, which is what Kommerce is, has a large addressable market and genuinely grinding unit economics, and the honest answer on winner-take-most is mostly no, because trust and logistics are stubbornly local and do not concede to whoever raised the most money. That answer should change how I finance it. Dressing a durable regional operator up as a global winner-take-all platform to clear a fund's billion-dollar bar would not make it one. It would only load a good business with a mandate to take risks it does not need to take, on someone else's clock.
The category error is not taking venture capital. It is taking it for a business whose best honest outcome the fund is structurally required to call a failure.