This is the single most useful piece of arithmetic in data center land, and it explains behaviour that otherwise looks irrational from the outside.
The comparison
Two 100-acre parcels, from an illustrative model published by DataCenterScope:
| Site A | Site B | |
|---|---|---|
| Price per acre | $1,200,000 | $400,000 |
| Land cost | $120M | $40M |
| Enabling infrastructure | $30M | $100M |
| Delay | — | 24 months |
| Timing cost at $1.5M/month | — | $36M |
| Effective total | $150M | $176M |
Site B saves $80 million on land, spends $70 million more on infrastructure, and loses $36 million to delay. Net: $26 million worse.
The three components
Land
The only one anybody quotes, and frequently the smallest.
Enabling infrastructure
Substation, transmission upgrades, fiber extension, water and sewer, road and access work, grading. On a constrained site this can exceed the land cost several times over. Who actually pays for the substation →
Timing
The one that gets missed. Financing cost on committed capital, deferred revenue, and contractual exposure to a customer expecting capacity on a date. The model prices it at $1.5 million per month; the real figure varies by project size and capital structure, but it is never zero and it is rarely small.
What this explains
Why developers pay enormous premiums for powered sites. They are buying out the second and third columns.
Why a landowner with slow power cannot get constrained-market pricing. The buyer has to fund the gap, and it comes out of the land price.
Why speed to power outranks everything. It is the only variable that moves two of the three columns at once.
For landowners
This is why "my land is worth $8 million an acre because that is what Virginia gets" does not hold. The Virginia number is the price of a site where columns two and three are already small. If your parcel needs a $70 million substation and a two-year wait, the buyer deducts that from what they will pay for the dirt. They have to — the project has one total budget.
The flip side is worth knowing too: anything that reduces the buyer's column two or three raises what your land is worth. Existing industrial zoning. An adjacent substation with headroom. Fiber already at the property line. A utility with large-load experience. These are not cosmetic. They are money.
The Site A / Site B comparison is an illustrative model published by DataCenterScope, not transaction data. Actual figures vary enormously by project.