Reframing the Constraint: From Deal-by-Deal to Portfolio Scarcity

Expert site selection begins with an uncomfortable admission: the skills that win a single deal can lose a portfolio. When you chase one building, you optimize locally. You find the cheapest power, the friendliest county, the fastest permit, and you sign. Do that eight times independently and you discover the trap. You have concentrated your load in one utility's territory, exhausted the goodwill of one community, and stacked your in-service dates so that they all land in the same overheated construction quarter. A portfolio is not a stack of deals. It is a single system with shared constraints, and the binding constraint is almost never the one that feels urgent in any given negotiation.

The reframe is this: stop asking "where is the best site?" and start asking "what is the scarcest thing I am allocating across sites, and how do I ration it?" For nearly every operator building at scale today, that scarce thing is entitled powered land. Break the phrase apart. "Land" is the easy part; raw acreage is plentiful and cheap almost everywhere you would not want to build. "Powered" means the parcel has a credible, contracted path to firm megawatts on a known date, not a hopeful letter from a utility. "Entitled" means the zoning, environmental clearances, water agreements, and local approvals are either in hand or on a de-risked path. A parcel that has all three is rare. A parcel that has two of three is a project waiting to stall.

Why is the combination so scarce? Each leg is governed by an institution that moves on its own clock and answers to its own constituency. Power sits behind the utility and its regional grid operator, whose interconnection queues in the busiest markets now stretch years and whose transformer and substation supply chains are genuinely constrained. Entitlement sits behind county commissions, planning boards, and state environmental agencies, each responsive to local politics that can turn against large loads overnight. Land assembly sits behind private owners, title complexity, and sometimes decades-old easements. The three institutions do not coordinate. They cannot be paralleled indefinitely. And crucially, progress on one does not transfer to another: a shovel-ready entitlement on a parcel with no power path is worth very little.

This is why capital is not the constraint. Money is fungible and, in the current cycle, abundant; infrastructure funds, hyperscalers, and sovereign vehicles are all competing to deploy. Silicon is a real constraint but a self-healing one, because fabs expand and allocation eventually loosens. Neither of those responds to the specific, local, non-transferable frictions that gate a powered entitled parcel. Faraday's read: the operators who will still be standing at the end of this build cycle are not the ones who raised the most or secured the most chips. They are the ones who quietly optioned entitled powered land two years before they needed it, while everyone else was still modeling cost per kilowatt-hour.

At portfolio scale, this reframe changes what you measure. Your primary inventory metric is not square footage under roof or capital deployed. It is contracted firm megawatts on entitled parcels, sequenced by in-service date. That is the number that tells you whether you can actually deliver the capacity you have promised to tenants or to your own compute roadmap. Everything else, including the choice between build-to-suit, powered shell, and owning versus leasing, is downstream of how much of this scarce asset you have secured and when it comes online.

The strategic implication is that acquisition and options should run ahead of demand, not in response to it. In a normal real estate business, building ahead of demand is speculative and dangerous. In a power-constrained data center portfolio, the danger runs the other way: if you wait until you have signed a tenant to start assembling entitled powered land, you have already lost, because the entitlement and interconnection calendar is longer than any tenant will wait. The expert operator carries a pipeline of optioned parcels at varying stages of entitlement and power maturity, treating that pipeline as the real balance sheet. Deals are drawn down from it. This is the mental model the rest of the course builds on: land inventory as reserve currency, sequenced deliberately, defended jealously.