AI infrastructure is becoming a regional development test

An AI campus arrives with a technology story. The local development questions are more familiar: where will the electricity come from, how will the facility be cooled, and who will pay for the infrastructure needed to make it work?

Those questions deserve attention before a community commits land, incentives or public investment. A large announced budget tells an economic development team very little about when a project can operate, how much utility capacity it will need in its first phase, or which costs sit outside the developer’s property line.

For infrastructure firms, the opportunity extends well beyond the data hall. It includes the electrical connections, civil works, cooling systems and operating arrangements that turn computing equipment into a usable facility. The difficult part is getting those pieces ready together.

Start with the operating requirements

Lawrence Berkeley National Laboratory’s 2025 update to its U.S. data center energy report, published in June 2026, estimates that data centers could consume 11.8% of U.S. electricity by 2030. Its wider scenario range is 9.5% to 15.3%. Those figures cover the data center industry as a whole, rather than AI facilities alone, and remain projections.

The national numbers explain why the sector attracts attention. A development team still needs a much more specific brief: the expected computing workload, electrical demand at opening, expansion plans and acceptable interruptions. Those assumptions should guide site selection and design.

AI training can also create a different electrical profile from a steady industrial load. In a May 2026 technical discussion, the Department of Energy described how synchronized computing activity can produce rapid fluctuations in power demand. Some can interfere with nearby generating equipment. DOE highlighted research into monitoring approaches that can capture those changes.

The practical implication is that a utility conversation should cover how a campus will operate, as well as how many megawatts it requests. Developers should bring their electrical designers and operating teams into that discussion early enough to influence the service plan.

Cooling belongs in the site decision

Water requirements need the same level of detail. “Liquid cooling” is a description of part of a system, not a complete account of a facility’s water demand.

DOE’s guidance on data center cooling water efficiency illustrates a liquid-cooled arrangement in which a closed loop carries heat from computing equipment to another loop connected to cooling towers. Recirculating coolant at the equipment therefore does not, by itself, establish that the whole facility operates without evaporative water use.

A useful site review asks for the proposed cooling configuration, expected annual consumption, peak-day needs and the source of that water. Reuse options, wastewater requirements and operating conditions during drought should be part of the discussion where relevant. The answer will depend on the design and location; a claim made about one campus should not be carried over to another.

Build a delivery plan beyond the property line

For an owner, the next step should be a shared schedule linking the buildings to the infrastructure serving them. It should identify responsibility for utility studies, off-site works, equipment procurement, permits and commissioning.

Phasing deserves particular scrutiny. If the first building opens before the rest of the campus, the team should establish which electrical and cooling systems must be complete for that opening and which can follow later. Contractors need those decisions to price and sequence the work. Utilities need them to understand how demand will develop.

Local officials should ask for an equally clear account of public obligations. Which road or utility improvements would the community fund? What commitments depend on later phases? What happens if the developer changes its schedule or builds less than initially proposed?

The economic development case should distinguish construction activity from permanent operations. Employment estimates, procurement opportunities and tax projections need an explanation of their assumptions and timing. A headline investment figure cannot do that work.

For engineering firms, contractors and equipment suppliers, this creates a concrete business-development task: identify the owner, the delivery team and the actual package being procured. An announcement is a starting point for research. A defined scope, funded phase and credible procurement timetable make it an opportunity.

AI infrastructure development will reward places that can answer these questions clearly. Before celebrating the size of a proposed campus, establish what its first operating phase requires and who is responsible for delivering it.

Infraday analysis based on the public research and technical guidance linked above. Updated September 27, 2026.