Two 300-acre parcels sit across the road from each other. Same soil series, same corn, same 345 kV line running down the section road. One is worth what farmland in that county is worth. The other has a buyer who has never once looked at the corn. The difference between them is not the land.
Powered land is real estate where the electricity equation is solved before anything vertical gets built: interconnection rights and a real queue position, substation and transmission capacity, entitlements for the actual use, and diverse fiber. Grading, pads, and halls come after. The industry named it as a class in 2025 and 2026 — Hines, CBRE Investment Management, Linklaters — and CBRE IM's framing is the one worth keeping: value is shifting upstream from the buildings to system access, meaning power, interconnection, and permits.
Speed-to-power, in CBRE IM's phrasing, now sits alongside location as a primary determinant of land value. That sentence does a lot of work. It means a parcel is not powered because a transmission line passes nearby. It is powered when a utility has studied the load, named who pays for the upgrades, and committed to a date you can underwrite.
the layers that have to be solved
Powered land is not one thing secured. It is five, plus an optional workaround — and any one of the five can be the reason a site never energizes.
| layer | what actually gets secured | why it gates the deal |
|---|---|---|
| Power deliverability | Interconnection agreement, queue position, substation and transmission capacity, phased load plan, cost allocation | Queues in key US markets are quoted in years, not months. Watts are the binding constraint, not acres |
| Land assembly | Contiguous acreage with room for pads, substation, stormwater, laydown, and setbacks | Fragmented parcels add execution risk and break phasing |
| Entitlements | Zoning, conditional use, environmental review, community approval | A growing share of failed projects die here. They die locally, not technically |
| Connectivity | Diverse fiber routes, backbone access, latency to carrier hotels | Table stakes. A single fiber path is a resilience failure, not a discount |
| Water and cooling | Basin availability, utility water contract, air versus evaporative cooling path | Decides the engineering and the politics at the same time |
| Optional: behind-the-meter | On-site generation, storage, ability to run parallel to or independent of the grid | A workaround when the grid queue stalls, not a bypass around it |
Notice the order. Power comes first because it is the constraint that cannot be bought out of sequence. You can re-plat a parcel, re-route fiber, and re-file a permit. You cannot conjure capacity on a circuit that does not have it.
not powered land: a map line
Say it plainly. A map line is not powered land.
A GIS overlay showing 345 kV within a mile is a hypothesis about the future. It says nothing about available capacity on that circuit, the cost of the upgrade or who pays it, the queue ahead of you, or whether the utility will commit to a delivery date inside your hold period. Assembled acreage beside a corridor with no utility conversation is inventory. Zoning with no watts behind it is a permit in a drawer.
This is not a rhetorical flourish. It is where basis gets destroyed — capital pays a power-adjacent price for a parcel whose power story is an assumption, then carries it through a hold period while the assumption is tested by an engineer who was never in the room.
power is a land attribute, not a data center product
Powered land is a subset of infrastructure land — parcels where a utility attribute, rather than frontage or visibility, decides the highest use. Power is that attribute for more buyers than the headlines suggest.
| buyer | why power is the search criterion |
|---|---|
| Data centers | Large, continuous, high-utilization load. The loudest buyer in the market, and the one setting land basis |
| Battery storage and generation | The interconnection is the asset. Land is the cheapest input |
| Advanced manufacturing | Chips, batteries, and food processing carry process loads that small industrial parks cannot serve |
| Fleet and truck charging | Megawatt-scale depots need service that freight-corridor land rarely has |
| Cold storage | Refrigeration is a continuous draw, and it is the reason a site pencils or does not |
| Equipment and outdoor storage yards | A quieter case today, growing as fleets electrify |
This matters for two reasons. If you underwrite powered land as a data center bet only, you inherit the AI capex cycle as your only exit. And if you are a landowner being told your ground is only valuable to hyperscalers, that is a negotiating position, not a fact.
why powered land is valuable
Because megawatts are scarcer than acres, and the scarcity is getting priced.
Hines Research estimates that roughly 20,000 acres of powered land sit under operational data centers worldwide today, and that another 40,000 acres — close to 2 billion square feet, just under three Manhattans — will be needed over five years to meet growth projections through 2030. The arithmetic behind those numbers is a planning ratio of 3 to 4 MW of capacity per acre, using 3.5 as the midpoint.
Treat all three figures as what they are: estimates from a manager with a thesis, useful as the scale of the search rather than a verified inventory. They are directionally supported by the capital behind them. Following Q1 2026 earnings, the big five listed hyperscalers — Amazon, Alphabet, Microsoft, Meta, and Oracle — raised guidance to a combined 2026 capex now tracking $740 to $760 billion, per CBRE IM. That is raised guidance, not a signed contract.
The honest counterweight belongs in the same paragraph. Nobody gets to underwrite a seven-year hold on a five-year forecast without a haircut. A position that only works if the capex curve holds all the way through is a bet on the curve, not on the land.
What that scarcity really means: the shortage is not dirt. The shortage is power-ready — and it is why an acre with a studied interconnect trades at a multiple of the identical acre without one.
the acre has two halves
A campus does not consume every acre it buys. A site is pads, a substation yard, stormwater, laydown, access, and setbacks — and then the rest of the parcel, which still drains to the same creek, still sits over the same aquifer, still carries the same field edges and cover it carried before the option was signed.
The trade calls the whole thing dirt. The parcel does not stop being a watershed because a spreadsheet calls it dirt.
That remainder is where the deal is usually decided anyway. Water is among the most common objections in contested siting fights, and roughly 43% of data centers sit in high water-stress areas, per S&P Global — which is why the cooling path is an entitlement question before it is an engineering one. Farmland identity, noise, and the night sky do the rest of the work in a hearing room. We wrote the full inventory of what gets said in those rooms in why communities oppose data centers, and the water depth sits in data centers drink water. ensurance refills the glass.
Here is the part conventional underwriting has no line for. Watersheds, working farmland, and living cover on and beside a power corridor exist whether or not anyone buys a certificate or options an interconnect. Ensurance is how the remainder and the same-basin systems get funded as the campus is assembled — not what those systems are. Insurance pays out after the damage; ensurance funds the living system up front, while it is still working.
In plain instrument terms: a certificate funds one named place and its living systems directly; a coin funds protection across the protocol more broadly. Both are ways to commit money to the remainder now, during the years the watts are still being negotiated, rather than promising it in a press release later.
And the limits, stated up front. We do not build hyperscale pads, energize sites, or move anyone's place in a queue — those sit with utilities, state commissions, FERC, and the ISOs, and anyone telling you otherwise is selling something. What we do is find, broker, option, and help entitle infrastructure land, and route the living remainder into funding that is committed in the present tense. Volumes today are small. The conventional side of the market buys the power option; this is the other half of the same acre.
frequently asked questions
what is powered land?
Powered land is real estate where the electricity equation is solved before vertical construction: interconnection rights, substation and transmission capacity, entitlements, and fiber secured upstream. It is valued on deliverable megawatts and timing, not on proximity to a line.
is powered land the same as a data center site?
No. Every hyperscale data center site is powered land, but not every piece of powered land becomes a data center. Powered land is a subset of infrastructure land — ground where electricity is the deciding attribute. Battery storage, advanced manufacturing, cold storage, and charging depots buy the same attribute. Data centers are the loudest buyer, not the only one.
why is powered land valuable?
Because interconnection capacity is scarcer than acreage, and the wait to get it is measured in years. A parcel with a studied load, a named cost allocation, and a committed energization date compresses the single longest item on a developer's schedule. That compression is what the premium buys — not the soil.
what to read next
If you are screening sites, the next post is the underwriting version of this one: a map line is not powered land separates strong inventory from speculative inventory. If you own ground near a substation and the phone has started ringing, start with the acre next to the substation.
See how this lands for infrastructure investors →
See how it lands for data center developers →