← The Index
005
Field Note · 005
The seat and the meter.
On the quiet rebalancing of software business models. Spring 2026.
The software industry has run on two pricing logics for a long time. The first is the seat: a license, an annual price, a login. The second is the meter: an API key, a price per call, a usage curve that surprises someone in finance every quarter. For two decades, the seat won. In the last eighteen months, the meter has been catching up.
01 /
The pejorative for the new generation of AI-native products is "wrapper." The implication is that the product is thin, that the moat is borrowed, that the real work is happening at the model provider. Some of this is true. Some of it is a category error. A SaaS application is a wrapper around a database. The interesting question is not whether something is a wrapper. It is whether the wrapping does work the model alone cannot do.
02 /
Seat pricing assumed engagement was good. The more you used the product, the more value you got from the seat you were paying for. Consumption pricing inverts this. The product owner wants the customer to get their job done in the fewest possible tokens. The customer wants the same thing. The user metrics that worked under SaaS now point the wrong direction.
03 /
Enterprise procurement does not yet know how to buy software whose price scales with usage no one can forecast. CFOs read the trailing month and adjust the next month's budget. Legal asks for a cap. Security wants to know what the bill looks like under an automated agent that does not sleep. None of these questions are about features. All of them are shaping how this category gets adopted.
The product category that wins this decade will not be defined by who has the best model or the cleanest UI. It will be defined by who teaches the buyer how to read the bill, and what to expect from a piece of software whose appetite is a variable.