Episode 2 · 30 Jun 2026

The price of a prompt

Per-seat, per-token, per-outcome: AI pricing models are arguments about who carries the risk. How to read a pricing page as a strategy document.

Show notes

• Every AI pricing model is a bet about where the cost goes. Per-seat bets on stability, per-token bets on efficiency, per-outcome bets on the model actually working.
• The pricing page is the most honest strategy document most vendors publish.
• When a vendor reprices, the question is never just how much. It is whose risk increased.

Chapters

• The week in five lines
• Three pricing models, three risk positions
• The pricing page as a strategy document
• Listener mail

Transcript

This episode runs as written notes rather than a recording. The argument, in order.

AI pricing looks like arithmetic and reads like strategy. Charge per seat and you are betting usage stays predictable. Charge per token and you are handing the customer the meter and betting on your own efficiency. Charge per outcome and you are making the boldest claim of all: that the model works well enough to be paid on results.

Each model moves risk to a different side of the table. That is why a pricing page is worth reading the way you would read a roadmap. It tells you what the vendor believes about its own costs, its own reliability and its own future discounts.

For buyers, the exercise is to price your own usage under each model before signing anything. The cheapest page in the demo is regularly the most expensive contract in production.