Subscribe
Analysis

The Inference Economics Nobody Modeled

Training got the headlines. Inference will rewrite the unit economics.

↳ Most pricing pages will be wrong by Q3. The companies that fix this first will own the margin.

Most published cost discussions about AI focus on the training run. That is a mistake. Training is a one-time capital expense. Inference is an operating expense that scales with usage, and at the unit volumes the next decade is heading toward, inference is where every interesting margin question lives.

I have been collecting unit economics from operators across the stack for the last six months. The picture that emerges is, frankly, not the one most pricing pages assume.

Most pricing pages will be wrong by Q3. The companies that fix this first will own the margin.

The full essay is a long piece of work and I am still finishing it. Subscribe and you will get it when it ships.

I write to think.
You can read along.

I'm Rijul. I write essays, host a podcast, and build small things on the web — all of it in service of one question: how do we leverage AI in the next decade without giving away what mattered in the last? New work lands here when it's ready. Subscribe and I'll send it once.