AI Ledger Hepburn Advisory Hepburn Advisory
Overview / AI Companies / Model providers

Eight labs publish a revenue figure. Two publish what it costs them to earn it.

The frontier labs are the most-discussed companies in the AI economy and among the least-disclosed. This page puts what is actually on the record for each one in a single place — run-rate revenue, collected revenue, operating loss, funding chain, compute commitments — and states plainly what is missing.

One block per company, in run-rate order. Every figure dated, graded and linked; every gap named.

Application layer Model providers Compute providers
— —

Tracked, no published revenue

How these profiles are built

Directional arithmetic, not a forecast. Every figure is derived at build time from dated claims held against each company; nothing on this page is typed into the markup.

  • A profile needs evidence, not just a number. A company earns a profile when its run-rate is backed by a dated, graded claim — a company disclosure or a named outlet. A figure sitting in our records with nothing behind it is not published here; the company appears in the roster instead.
  • Run-rate is not revenue. Run-rate annualises current sales. Collected revenue is what was billed over a year. Where both are held, both are shown, because the gap between them is the growth rate.
  • Statutory loss never appears alone. Where a headline loss is dominated by a one-off non-cash accounting item, the operating loss is rendered beside it in the same block. Neither figure can be read without the other.
  • Projections are excluded. Forward-looking years held in our records are never rendered as current figures.
  • What is missing is stated. Each profile ends with what we do not hold. An absent funding history is a fact about the record, not a zero.

Prepared by Hepburn Advisory. Last updated —.