The AI economy, in five questions.
The Ledger tracks the build-out from capital committed to revenue collected, compute monetised, usage consumed and power required. Each ledger answers one question, with the evidence underneath.
Get a short brief when the numbers move.
Hepburn Advisory emails a short brief when a headline figure or the method behind it changes. One list covers Ledger updates and Hepburn news.
Subscribe via Hepburn AdvisoryHow much has been committed before revenue catches up?
Infrastructure is a stock that now has to earn through a flow of customer revenue.
What money has actually arrived from customers?
Customer-paid revenue is the floor against which other claims are reconciled.
Who is monetising the build-out now?
Hyperscalers and neoclouds are already earning from the supply side.
Is demand real, paid, or subsidised?
Token activity is large. Monetisation quality is uneven.
Can the revenue curve be physically served?
Capacity, not only capital, determines how much of the curve can be delivered.
Every layer against the $1 customers paid.
Each layer’s 2026 flow against the same customer-paid revenue baseline, with 2025 alongside it. Every layer is bigger than it was — the multiples are smaller because revenue is growing faster than any of them. Capital keeps its separate cumulative stock lens.
Basis: 2026 annual flow · 2025 comparator
The questions this raises
Three open questions the reconciliation puts on the table.
How much of the infrastructure already built can be covered by customer-paid revenue, and how quickly?
Which suppliers can keep serving buyers if free usage, subsidy or resale margins tighten?
Where does the system bind first: capital, compute supply, token demand or power?
What moved this month
Get a short brief when the numbers move.
Hepburn Advisory emails a short brief when a headline figure or the method behind it changes. One list covers Ledger updates and Hepburn news.
Subscribe via Hepburn AdvisoryHepburn Advisory
The AI Ledger is published by Hepburn Advisory, the independent practice of Simon Bowker. Hepburn works with AI founders on go-to-market and with enterprise leaders on what AI costs and what it returns.
Contact Hepburn AdvisoryMethodology caveat
Projections are directional arithmetic, not forecasts. Benchmarks are dated and sourced. Every ratio on this page defaults to the basis named beside it; every assumption is editable. Where a number appears on a Ledger page, that page is the canonical source. Full method on the Methodology page.