The power law is real. It’s also a choice.

The most quoted dataset in venture capital comes from Horsley Bridge, an institutional investor with stakes in funds that backed roughly 7,000 startups between 1985 and 2014. The finding has become scripture: about 5% of the capital deployed generated 60% of all the returns. A tiny fraction of the investments produced the overwhelming majority of the gains. This is the power law, and it governs how most of the venture industry actually behaves, whatever it says in its pitch decks.

The standard response to that data is portfolio math, and on its own terms it’s rational. If only a handful of investments will ever matter, then the optimal move is to buy as many tickets as possible, maximize your exposure to the extreme outcome, and accept — plan for, even — that most of your companies will go to zero. Spray capital across enough long shots and you improve your odds of owning one of the few that returns the entire fund. For a manager paid on the aggregate performance of a large portfolio, this is the correct strategy. The arithmetic is sound.

It is also, viewed from the other side of the table, quietly brutal. The model only works because the manager has priced 95% of their founders at zero before those founders have written a line of code. Those founders don’t know it. They take the meeting, sign the term sheet, and believe they’ve found a partner — when in fact they’ve been bought as one lottery ticket in a book of hundreds, and the moment they stop looking like the one-in-a-hundred outcome, the attention evaporates. The model never required them to succeed. It only required one of their cohort to.

SwellPoint is built on a deliberately different response to the same data. We invest in small companies that already have real revenue, real customers, and real unit economics — businesses where the base case is a company that simply works, not a company that died in an interesting way. We don’t need any single investment to return the entire fund, because we never priced the rest of the portfolio at zero to begin with. Every company we back is underwritten to be a company that lasts, on its own merits, whether or not it ever becomes a headline.

Founders deserve an investor whose model requires them to succeed.

Does that cap our theoretical upside? Measured against a true lottery-ticket portfolio, maybe it does — we are unlikely to own the one company in ten thousand that defines a decade. We’re at peace with that, because we think the power law is widely misread. The power law describes venture portfolios; it does not describe businesses. A well-run small company compounding steadily for a decade is not a tail event or a statistical miracle. In the survival data it’s simply a survivor — and in our underwriting it’s the plan, not the fantasy.

There’s a subtler point hiding in the Horsley Bridge numbers that the lottery-ticket crowd skips past. The concentration of returns is partly a description of reality and partly a self-fulfilling prophecy of the strategy itself. If you fund a hundred companies and only have the bandwidth to genuinely help three of them, you have guaranteed that the other ninety-seven underperform their potential — and then you cite their underperformance as proof that the power law was inevitable. Attention is a capital input. Spread it to zero and you manufacture the very distribution you claim to have merely observed.

So we do the opposite, on purpose. We take fewer positions, we hold them longer, and we work them harder — bringing operating help, not just a wire transfer, to every company in the book. It means our portfolio will never have a hundred names in it. It also means no founder we back is secretly priced at zero. Every company we invest in is meant to matter, and is resourced as though it does. That isn’t a marketing line about being founder-friendly. It’s the underwriting itself — the actual model by which we expect to make money.