The power law

Venture returns do not look like a bell curve. They look like a power law: most investments return little or nothing, a thin middle returns a few times the money, and a tiny tail returns so much that it pays for everything else. Once you see that shape, partner behavior stops looking greedy or random and starts looking like optimization under a hard constraint.

Paul Graham called the strategy black swan farming — hunting outliers rather than maximizing the hit rate of pretty-good companies. The rest of this section is the arithmetic that makes the farming rational, and the misalignment it creates with founders who would be happy with a solid mid-size exit.

When you evaluate equity, ask whether the company's story to investors is power-law shaped. If the pitch is "profitable regional services firm," a VC cap table is a smell. If the pitch is "infrastructure for every AI agent," you are in the farming business.

Concepts in this section

Step	Concept	What it answers
1	Power-law returns	What does the distribution of deal outcomes actually look like?
2	Fund-returner math	How big must one exit be to return the fund?
3	Black swan farming	Why do partners prefer huge markets and high failure rates?
4	Founder and VC misalignment	Why a $50M sale can delight you and bore them.

Read Paul Graham's essay once, then re-read the $50M vs $1B contrast until it is automatic.

Black Swan Farming (Paul Graham)Black Swan Farmingpaulgraham.com

Common mistakes

Raising venture money for a calm, profitable niche "to see what happens." You inherit black-swan board expectations.
Judging VCs by hit rate. They cannot optimize that after fees.
Ignoring ownership at exit. A tiny stake in a unicorn does not return a large fund.

Key takeaways

Venture returns follow a power law: a few deals produce most of the dollars.
Funds farm outliers; hit rate is not the objective.
At 10% ownership, a single company often needs a ~$1B+ exit to return a $100M fund alone.
A founder-friendly $50M exit can be irrelevant to fund DPI.
If your upside is capped or control matters more than speed, prefer non-VC paths.