Valuation Mechanics: Cap Rates, Yield-on-Cost, and the Powered-Shell Arbitrage
Data-center M&A is priced with the tools of infrastructure finance, not the intuition of real estate. Three numbers do most of the work: the capitalization rate, the yield-on-cost, and the spread between them. Master those three and you can reconstruct the logic of nearly any deal in the sector, because every developer and fund is running the same arithmetic.
Start with the capitalization rate, or cap rate. It is simply the annual net income of a stabilized asset divided by its price — the yield a buyer accepts to own contracted cash flow. A low cap rate means a high price for a given income, and cap rates fall as cash flow becomes more contracted, longer-dated, and backed by stronger credit. A speculative asset trades at a high cap rate because the income is uncertain; a fully-leased asset with a top hyperscaler tenant trades at a low cap rate because the income is bond-like. The cap rate is therefore a direct readout of risk: it compresses exactly as the lease strengthens. This is the same repricing described earlier, now expressed as the number the market actually quotes.
Next, the yield-on-cost. This is the annual net income divided by the total cost to build the asset — land, power, construction, and everything else — rather than by its market price. It measures the return a developer earns for creating the asset from scratch. A developer who can build for a given cost and contract a given rent earns that yield-on-cost as a development return. Crucially, yield-on-cost is usually higher than the cap rate a stabilized asset commands, because building carries execution risk that the market rewards, while buying a finished, contracted asset does not.
The gap between these two numbers is the entire profit engine of the sector, and it has a name worth remembering: the development spread, or the powered-shell arbitrage. Here is the mechanism in full. A developer creates a powered shell and builds it at a cost that implies, say, a high yield-on-cost once contracted. It then signs a long-term hyperscaler lease. That signature transforms the asset into bond-like cash flow, which the market values at a low cap rate. The developer now sells — or refinances — the asset at a price set by that low cap rate, well above its build cost. The difference between the low cap rate the market pays and the high yield-on-cost the developer built at is captured as profit. Create cheaply, contract, sell dear. That is the arbitrage, and it is the reason capital floods into development.
Work the direction of each lever. Anything that raises the yield-on-cost — cheaper power, faster construction, land bought before a metro became strategic — widens the spread. Anything that lowers the exit cap rate — a longer lease, a stronger tenant, fixed escalators, a scaled platform that finances cheaply — also widens the spread. The developer's whole craft is pushing those two numbers apart: build for a higher yield-on-cost, exit at a lower cap rate. The wider the gap, the greater the value manufactured per project.
This arithmetic explains behavior up and down the market. It explains why developers race to secure cheap power and land ahead of demand: they are buying a high yield-on-cost before the market catches up. It explains why the hyperscaler lease is the pivotal event: it is the signature that collapses the exit cap rate and crystallizes the spread. It explains why funds pay low cap rates for stabilized assets: they are buying durable, cheaply-financeable cash flow, and their low cost of capital lets them accept a low yield. And it explains the rotation of the sector — developers manufacture spread, funds buy the stabilized result, capital recycles into the next build.
The reader can use these three numbers as a diagnostic on any transaction. Given a deal price and the asset's income, back out the cap rate; a low one signals strong contracted cash flow, a high one signals speculation or weak tenancy. Compare it to the plausible yield-on-cost of building the same asset; the gap tells you how much development spread is on the table and why a developer would want to create rather than buy. When a headline multiple looks extreme in either direction, these numbers usually resolve the puzzle. Valuation in this sector is not mysterious. It is the disciplined application of cap rate, yield-on-cost, and the spread between them — and the spread, opened by cheap creation and closed by a hyperscaler signature, is where the money is made.