Underwriting the Power Law

Every asset class trains its investors to think in averages — expected returns, downside protection, batting averages. Venture punishes all of it. Returns in early-stage portfolios follow a power law: a tiny number of outcomes generate the majority of all value, the median investment loses money, and the difference between a good fund and a great one is usually a single company. Until this distribution is internalized — not acknowledged, internalized — every portfolio decision will quietly fight it.

What the distribution actually implies

In a typical seed portfolio, roughly half the companies return little or nothing, a band returns capital or a modest multiple, and one or two drive the fund. The implications are brutal and liberating in equal measure:

• Your losers are irrelevant to fund outcome. A written-off seed check costs its check size — bounded, survivable, already priced into the model. Obsessing over loss prevention optimizes the part of the distribution that doesn't matter.
• Your winners are everything, and capped upside is the true catastrophe. The unbounded right tail is the entire source of venture returns. Anything that truncates it — premature exits, structure that caps outcomes, ownership too small to matter — is worse than a loss, because it spends a slot in the portfolio without buying tail exposure.

The fund-returner test

The discipline that operationalizes the power law is one question, asked of every deal: can this check plausibly return the fund?

The math is unforgiving. Required outcome equals fund size divided by realistic ownership at exit. A $50M fund holding 5% at exit needs a $1B outcome for one fund-returner — and needs it after dilution, which means double-digit entry ownership or aggressive follow-on defense. Run your own numbers once and the conclusions write themselves: small checks into large rounds cannot return your fund no matter how well the company does, and "good" outcomes — the $80M acquisition everyone celebrates — are rounding errors at the fund level unless your ownership is exceptional.

This test is not pessimism about modest outcomes. It is the recognition that a portfolio slot spent on a company with a structurally capped ceiling is a slot not spent buying tail exposure — and tail exposure is the product.

Underwriting upside, not survival

Traditional diligence asks "what could go wrong?" Power-law diligence asks the stranger question: "if this works, how big is it — and what has to be true?" The memo discipline that follows:

• Write the specific, falsifiable path to a fund-returning outcome: market, ownership, dilution, exit size. If constructing that path requires historic best-case everything, the deal fails the test at any price.
• Give weirdness a fair hearing. The fund-returners of every vintage looked overpriced, too early, or faintly absurd at entry — consensus deals are consensus-priced, and consensus pricing has no tail in it. This is not a license for credulity; it is a reminder that "obviously good" and "capable of 100x" are nearly disjoint sets.
• Let the bear case be about the ceiling, not just the floor. "This survives but tops out at $200M" is a pass for a fund whose math needs a billion.

Where investors fight the math

The power law is easy to quote and hard to obey. The recurring failure modes:

• Selling winners early. Taking 3x off the table in the company tracking toward 50x is the most expensive risk management in finance. Partial secondaries have their place; reflexive derisking of the right tail does not.
• Reserving into weakness. Follow-on capital drifting to struggling companies — where allocation is plentiful — instead of defending ownership in compounding ones, where it's scarce.
• Diversifying past the point of relevance. Sixty positions at 1-2% ownership is index construction with venture fees; the tail can land in your portfolio and still not return your fund.
• Structure-brain. Importing downside protections that cap or complicate upside imports another asset class's physics into one where they don't apply.

The temperament it demands

Power-law investing means being wrong, visibly, most of the time — and being emotionally capable of concentrating support behind the one time you're spectacularly right. The score that matters is not batting average but slugging: not how often you're right, but how much it pays when you are. Build the portfolio, the reserves, and the memo discipline around that single asymmetry, and the distribution stops being a threat and starts being the strategy.