The Power Law: Why Most Startup Bets Lose, and You Still Win

Someone new to startup investing usually carries an instinct from every other asset they own: spread the money, avoid losers, aim for a steady average. Applied to startups, that instinct quietly destroys returns. Startup investing does not pay you for a good batting average. It pays you, almost entirely, for the one or two bets that go enormously right, and learning to feel comfortable with that fact is the real beginning of investing well.

What the power law actually says

In most asset classes, returns cluster around an average and extreme outcomes are rare. Startup returns do the opposite: they follow a power law, where a small number of investments account for nearly all of the gains and the rest return little or nothing. AngelList's data on thousands of early-stage investments shows this clearly, and the same shape appears in fund data: across a large sample studied by Horsley Bridge, a small share of deals produced the large majority of the returns. The distribution is not a flaw in the asset class. It is the asset class.

One calibration is worth adding. The most extreme version of this picture, dozens of bets to find a single hundred-fold or thousand-fold winner, comes from the largest US venture markets. In Taiwan the return distribution tends to be flatter, with outliers landing more often in the rough range of twenty to thirty times rather than a hundred or more, and local practice leans toward a more boutique, selective style of sourcing. The power-law logic still holds: a few bets still carry the result, and missing the winner still hurts most. The slope is simply gentler, so the same principle applies with somewhat less extremity.

Why this inverts your instincts

Two habits that serve you everywhere else become dangerous here. The first is loss-avoidance. If you screen primarily to avoid losing money on any single deal, you will systematically pass on the wild, polarising companies that have the highest chance of being the outlier, because those are exactly the ones that look riskiest at the seed stage. In a power-law world the cost of a loss is capped at one times your cheque, while the cost of missing a winner is unbounded. You are not protecting the portfolio by avoiding losers. You are starving it.

The second habit is optimising for the average deal. A new investor often tunes their screen to raise the quality of the typical company they back. But the typical company barely matters to the result; the outlier is everything. Sequoia and the best early-stage firms organise around this openly: they are not trying to be right often, they are trying to be enormously right occasionally, and they accept a high rate of zeros as the price of access to the few companies that return the fund. The mental shift is from 'will this lose money' to 'if this works, is it big enough to matter'.

What the power law does not excuse

It would be easy to misread all this as 'spray money everywhere and hope.' The power law rewards outlier exposure, but it does not reward carelessness, and the two questions still have to be answered together. You raise your odds of touching an outlier in two ways at once: by taking enough quality shots that you are statistically likely to be in one, and by keeping the bar high enough on demand and team that each shot has real outlier potential. A disciplined screen and a power-law mindset are partners, not opposites. The screen decides which companies could be the outlier; the power law decides how to size and spread your bets across them.

That is why this idea comes second in the series, right after the first-pass screen. The screen keeps your quality high on each individual bet; the power law tells you how many such bets to make and how to think about the ones that fail. The next piece turns this into concrete portfolio maths, how many companies, how much per cheque, and how much to hold back for the winners.

The takeaway

The single hardest adjustment for a new startup investor is emotional, not analytical: most of your bets will not work, and that is not failure, it is the design. Your job is not to avoid losses but to make sure you have real exposure to the rare company that returns everything, and to have the discipline that each bet could plausibly be that company. Judge any single investment by whether it could be the outlier, and judge yourself by the portfolio, never by the bets that went to zero along the way.

It is the logic every serious venture investor runs on, EvoScale Capital included: back a focused set of B2B companies that each have a credible path to being the outlier, and let the power law do what it does. The discipline this series teaches is how to make each of those bets a good one. The power law is why you make more than one.