Skip to the lesson

Module 1 · Review

The Capital-Formation Thesis

Faraday’s take

AI infrastructure is a capital-formation story before it is a technology story. The building costs run to hundreds of millions before a chip is installed, and no operating cash flow covers that up front. So the binding constraint on who builds capacity is not chips or talent but access to debt: who can assemble the money, at what cost, in what structure. A leased data center behaves like a bond, which lets it carry heavy borrowing at low rates and shrink the scarce equity check. Whoever commands the cheapest cost of capital can bid more for land and power, sign tenants at tighter rents, and still clear their hurdle. The capital stack, not the silicon, draws the map of where AI gets built.

Knowledge check

Why is a signed hyperscaler lease so valuable to a lender financing a data center?

Show the answer

Because it converts the building into a stream of contracted, predictable cash flow backed by a creditworthy tenant. The lender is effectively underwriting the tenant's long-dated promise to pay, which is what services the debt, so the loan prices off the tenant's credit rather than the bare concrete.