The quiet economics of leverage

The four forms of leverage are well-known. Labour, capital, code, media. The prices they actually charge — time, judgement, reputation, sovereignty — are not.

The standard taxonomy

Everyone who has read enough Naval can recite it: labour and capital are the old forms; code and media are the new. The new ones have zero marginal cost of replication. The old ones don't. The new ones compound while you sleep.

This is all true. It is also where most founders stop reading.

What each form actually costs

• Labour costs you management capacity. Not money — time and focus. The second employee is not twice the cost of the first; it is roughly three times, because coordination grows non-linearly.
• Capital costs you optionality. Money taken is money with terms. Every round is a narrowing of the set of futures you can pursue without explaining yourself.
• Code costs you maintenance. The line you shipped today is a liability you own. Unless it pays rent proportional to its complexity, it is a slow leak.
• Media costs you sovereignty. An audience you built watches you. You now owe them a version of yourself. Some founders underwrite this cheerfully. Most underwrite it without realising they have signed.

The founders I respect most are the ones who can name the bill for every form of leverage they are using at any moment.

The arbitrage

The arbitrage is not between the forms. It is between what each form charges you and what it charges the market.

Capital is dear to you but cheap to the market right now. Media is dear to the market but occasionally nearly free to build. Code, in 2026, is becoming one of the strangest forms of leverage to price — because the cost to produce has cratered but the cost to maintain has not.

The instrument you compound on should not be the one that is cheapest to set up. It should be the one that is cheapest for you to hold over a decade.

Pick leverage by its holding cost, not its construction cost.

The founder-scale view

At the single-operator scale, the interaction is unusual. You are the bottleneck for every form. You can hold more capital than labour because capital does not require management. You can hold more code than capital if you are technically literate. You can hold almost unlimited media — but only until the audience starts demanding that you be a person, at which point the bill arrives.

The final leverage — the one nobody puts on the list — is institutional trust. This one has the highest holding cost and the highest holding value. It takes a decade to build and an afternoon to spend. Most of the writing in this doctrine series, eventually, is about that form.

---

The honest lesson from running a portfolio for a decade: the leverage that looks most exciting on a pitch deck is usually the most expensive to hold. The leverage that looks boring — a mailing list that has survived three platforms, a reputation for returning emails — is usually the one worth doubling.