Put the national forecast in context
Berkeley Lab’s June 2026 update models data centres at 11.8% of US electricity use in its 2030 reference case. Its combined uncertainty scenarios span 9.5–15.3%. These are national model estimates, driven by equipment and operating assumptions; they are not a utility queue, inventory of available sites or a delivery commitment. Our screening interpretation is to compare each exact parcel, serving utility, phased load and required date against dated project evidence.
National forecasts provide context. They do not establish project-specific capacity or an in-service date.
Compare like-for-like enquiries
Use the same critical IT load, phased ramp, density and required date for each territory. Establish the exact parcel and serving utility before making a site-level statement. Regional activity does not establish whether a particular location can receive the requested service.
Include Atlanta and Ohio on the same basis
The sample map includes Northern Virginia, Dallas–Fort Worth, Phoenix, Atlanta and Ohio as research territories. It intentionally offers no live inventory or queue estimates. Apply the same utility, network, cooling, planning and commercial questions to each so the comparison reflects evidence quality rather than familiar market names.
Expose geographic gaps
A blank source field is a research task. Do not fill it with a neighbouring project’s experience or an old regional announcement. Record source age and the question the document actually answers. Refresh priority should follow the decision date and the materiality of the unresolved dependency.
Take this checklist into the next meeting
- Use a consistent load and date brief.
- Confirm parcel-level service territory.
- Request dated site-specific evidence.
- Record unknowns and source age without inventing rankings.
Evidence and assumptions
This guide develops the supplied buyer-side research framework. It makes no site-specific capacity claim. Obtain current utility documents, manufacturer specifications and appointed professional advice for the actual project. Any numerical examples are worked examples, not project data.