Methodology
How we screen land for data centers
One question runs through every engagement: can this land carry a data center, at what size, and what would it take? This page explains how we answer it, from a national screen to a single parcel.
Parcels screened
17.1 million
Largest single engagement
Nearly 500,000 acres screened in a single engagement for one of the largest landowners in the United States.
The stages of a screen
01
Scope the question
Every engagement starts with the decision it has to support: a target load and timeline, a region, a portfolio to triage or a single parcel to check. The load sets the acreage, voltage and utility needs everything else is measured against.
02
Screen
We run parcels through constraint layers, from national down to a single tract. Hard knockouts remove land that cannot work, such as cores in the floodplain or no realistic path to transmission. Graded factors then rank what remains by the size of data center each site could support.
03
Verify
Data shows where infrastructure is, not what it can do. For the short list we check what the maps cannot: available capacity, the utility’s position and timing, access, conditions on the ground and anything that would surprise a buyer later.
04
Assemble
Where a site needs more acreage, better shape or frontage, we identify adjacent parcels that work together and test the combined site against the same constraints.
05
Document
Each result comes with the reasoning behind it: what passed, what failed, the distances and sources, and what it would take to make the site work, in a form a developer, investor or board can check.
The constraint layers
Every site is checked against the same layers. Public data is where a screen starts, not where it ends: it shows where infrastructure and constraints are, and verification establishes what they mean for a specific site. The sources below are the kinds of data each layer draws on.
| Layer | What we look at | Where the data starts |
|---|---|---|
| Transmission and substations | Line voltage, distance to lines and substations, likely capacity and upgrade paths | Federal infrastructure datasets, utility and grid operator planning documents |
| Interconnection | Utility territory, large-load process, study status and realistic timing | Utility and grid operator filings, direct utility engagement on short-listed sites |
| Natural gas | Pipeline proximity, size and the feasibility of a lateral | Federal pipeline mapping, operator information |
| Water | Supply, discharge options and local constraints | Federal and state water data, local utility information |
| Fiber | Long-haul and metro routes, diversity and distance to network hubs | Carrier route information and network datasets |
| Floodplain | Mapped flood zones and base flood elevations | FEMA flood hazard mapping |
| Wetlands and waters | Mapped wetlands, streams and likely jurisdictional areas | National wetlands mapping, hydrography data, field review |
| Terrain and soils | Slope, elevation, soil types and geotechnical flags | USGS elevation data, USDA soil surveys |
| Zoning and land use | Current zoning, likely approval path, local restrictions | County and municipal records |
| Airspace and hazards | Airport proximity, height limits, natural hazard exposure | FAA data, federal hazard mapping |
| Access and shape | Road access, frontage, contiguity and buildable shape | Parcel data, road networks, aerial imagery |
| Incentives | Available tax programs and qualifying criteria | State and local program information |
Knockouts, then ranking
A screen works in two passes. Knockouts remove land that cannot work for the target load. Graded factors then rank what is left, so effort goes to the few sites with real potential. Sites are sized against the load they could support, not just a pass or fail.
Read more in our guides on screening a land portfolio and why sites fail.
| Tier | Load | Typical land |
|---|---|---|
| Edge | 1–10 MW | 1–10 acres |
| Enterprise & colocation | 10–60 MW | 10–75 acres |
| Hyperscale campus | 100–500 MW | 150–600 acres |
| AI & gigawatt campus | 500 MW – 1 GW+ | 500–2,000+ acres |
What comes back
Results are delivered in whatever form the decision needs, from a two-page answer to a full diligence file or live tools. Everything you share with us about land and projects is treated as confidential.
2 pp
Two-page brief
The answer, fast: does it fit, at what size, and why.
90+ pp
Full diligence report
Every constraint, distance and source, with the data behind each call.
PDF
Board-ready PDFs
Clean, cited documents for investment committees, lenders and partners.
Map
Online, searchable maps
Every parcel, layer and red flag on a live map your team can explore.
Tool
Live search tools
Filter and re-rank parcels yourself as your criteria or load changes.
App
Private web apps
A purpose-built app for your portfolio, pipeline or site search.
AI
An AI agent on your maps
Ask in plain English and get ranked parcels back, with the reasons.
Data
Data exports
The underlying results in the format your team already uses.
What a screen can and can’t tell you
A screen gives clarity on where to spend time and money, and why. It narrows thousands of parcels to the few worth pursuing and explains what each would need.
It does not replace utility studies, engineering, surveys, environmental work or legal review. Those confirm a site; the screen tells you which sites deserve them. Where a fact depends on a utility or agency, we say so.