Valuation is a claim on future extraction

Valuation is a claim on future extraction

A company is not valued because it is useful. It is valued because investors expect a stream of claims on future monetary output.

The first distinction is elementary but frequently lost in technology narratives. Enterprise value is a claim on operating cash flows available to debt and equity capital providers. Equity value is enterprise value plus cash and non-operating assets, less debt and other senior claims. It is not a measure of the number of people touched by a technology, the social surplus it creates, the replacement cost of all labor it influences, or the amount of GDP that passes through systems built with it.

EV = Σₜ unlevered FCFₜ / (1 + WACC)ᵗ

In a conventional discounted-cash-flow framework, enterprise value is the present value of future unlevered free cash flow discounted at the weighted average cost of capital (WACC). A multiple is merely a compressed expression of assumptions about those cash flows: growth, margins, reinvestment, risk, competitive duration and terminal economics.

This matters because AI narratives often move through an invalid shortcut:

Large economic impact → large TAM → large revenue → large valuation

Each arrow is conditional. A technology can affect an activity without becoming the seller of that activity. It can eliminate expenditure rather than inherit it. It can create customer surplus while capturing little producer surplus. It can produce a large temporary rent that competition rapidly compresses. Or it can create a durable bottleneck and justify extraordinary value. The thesis is therefore a burden-of-proof exercise: identify which arrow is being assumed and ask what preserves it.

A normalized enterprise value of 100 units is particularly useful because it forces scale into view. At a 25× revenue multiple, 100 units of value correspond to 4 units of annual revenue. At a 10× revenue multiple, it corresponds to 10 units of annual revenue. At a 25× free-cash-flow multiple, it corresponds to 4 units of annual free cash flow. These are not impossible numbers. But they are recurring monetary claims, not estimates of social usefulness.

The rest of the paper asks what kind of economic structure can sustain such claims when intelligence itself lowers the cost of creating, coordinating and substituting the things being sold.