The Data Center Map Gets More Selective
For EdgeCore, finding viable development locations has become an exercise in aggressive filtering. Kestler said the company evaluated 172 sites during the previous 12 months to narrow the field to seven locations it wanted to actively manage.
Its requirements include roughly 100 acres or more, the ability to support a 300-MVA-or-larger substation, credible utility development timelines and sufficient network proximity to support what Kestler called “interdependent compute” locations.
The distinction matters. Not every AI workload needs the same geography, and not every site marketed as available for AI infrastructure can support the combination of land, network, power and timing required to make a project real.
Miller placed that process in the context of a data center map already being redrawn by power availability.
Northern Virginia’s power constraints in 2022 provided an early warning, redirecting capacity into markets including Atlanta and driving developers farther afield in search of large blocks of electricity. AI has intensified the process.
As campus requirements move toward hundreds of megawatts and, in some cases, gigawatt scale, Miller said, fewer locations can satisfy all of the requirements simultaneously. Community acceptance is narrowing the map further.
At the same time, Miller pointed to a potential countertrend: the growth of AI inference could create another layer of data center geography. Some inference architectures may favor smaller, distributed facilities rather than concentrating every workload inside enormous campuses.
The result could be a more stratified infrastructure market. “Everything everywhere all at once,” Miller said.
Build Where Data Centers Are Wanted
For large campus development, Kestler offered another increasingly important filter.
EdgeCore wants to build where it is wanted.
In practical terms, that means targeting municipalities and jurisdictions that have already made deliberate decisions about where data center or other light industrial development belongs.
Kestler pointed to technology development zones and planned industrial areas in markets including Culpeper and Louisa County, Virginia, and Mesa, Arizona.
The goal is not to eliminate community engagement. It is to begin development in places where government, infrastructure planning and land-use expectations already recognize what is intended for the site.
Even then, community relationships require continued investment.
Kestler recalled earlier developments in Northern Virginia where residential neighborhoods arrived after data center facilities had already been built. That did not remove the operator’s responsibility to work with homeowner groups and address local concerns.
“We want to make sure that when we do go into a place, we’re a member of the community, not just a tax revenue base,” Kestler said.
Miller agreed that long-term alignment increasingly needs to begin much earlier, with municipalities clearly defining where development makes sense and developers researching the communities they intend to enter.
The same principle applies to utilities.
What Does It Mean to Have Power?
One of the most consequential exchanges of the opening session came when Vincent asked a deceptively simple question:
When a developer says a site “has power,” what must actually be true?
For Kestler, the answer goes well beyond a large number attached to a parcel of land.
A credible powered site should have site-plan approval that accounts for the substation, equipment on order, a completed load-impact study establishing a realistic service range and a load ramp agreed upon against the utility system’s ability to deliver.
Without those elements, large power claims deserve scrutiny.
The industry is entering a period, Kestler said, when “accuracy and deliberacy and transparency” around power and land will matter more.
That also changes the developer-utility relationship.
Kestler, who spent time at NextEra Energy after leaving Vantage Data Centers, said developers need to understand how utilities operate rather than entering meetings focused exclusively on what the data center needs.
A productive relationship begins with realistic discussions about timing, system capabilities, capital obligations and who pays for infrastructure at different stages.
His advice was deliberately simple: “Talk to them like they’re people.”
Getting More From the Grid We Already Have
Miller expanded the power discussion beyond traditional utility service.
Until recently, many data center operators viewed uninterrupted utility power as essentially non-negotiable. Grid constraints are forcing a closer look at flexibility.
The electric grid is built to accommodate peak demand, including the relatively small number of extreme periods when heat waves and other conditions push usage toward system limits. During much of the rest of the year, additional capacity may exist but remain difficult for large new loads to access under traditional arrangements.
That is creating new interest in batteries, curtailment agreements, workload management software and other strategies that can make data centers more responsive to grid conditions.
A facility able to ride through a critical period on battery storage, generators or flexible workloads may be easier for a utility to accommodate earlier than a load demanding maximum grid capacity around the clock.
Miller also pointed to emerging software and grid-analysis approaches designed to identify pockets of available power that conventional interconnection processes may overlook.
Behind-the-meter generation will remain part of the discussion as well.
But before assuming massive amounts of new generation and transmission can arrive quickly, Miller argued, the industry has a strong incentive to extract more usable capacity from infrastructure that already exists.
“Getting the most out of the grid we have,” he said, is likely to become increasingly important over the next several years.
Workforce Becomes Part of Site Selection
Power and land are not the only resources becoming harder to secure. Asked what keeps him up at night, Kestler pointed to labor.
The challenge is not simply finding people willing to work in data centers. New electrical architectures, higher voltages and changing equipment are creating training and safety requirements that evolve with the technology.
For EdgeCore, workforce availability now factors directly into development planning.
The company has invested in community college programs, technical education and apprenticeships in markets where it operates. Kestler pointed to graduates from technical programs in Culpeper moving into jobs across the broader digital infrastructure ecosystem, including electrical, plumbing, safety and other skilled trades.
The objective is to develop a durable labor pool capable of supporting campuses that may continue expanding for decades.
Miller said newer data center markets can learn from Northern Virginia by investing early in those education pipelines, particularly community colleges, high schools and the trades.
Power availability alone does not create a sustainable data center market if the workers needed to build and operate the infrastructure cannot be found locally.
Miller also cautioned against allowing public discussion of artificial intelligence to obscure the broader role of the infrastructure.
“AI does not equal data centers,” he said.
Data centers support AI, but they also support streaming, communications, cloud services and the wider internet economy on which consumers already depend.
That distinction may become increasingly important as the industry competes for workers and community support at the same time.
Building for Hardware That Hasn’t Arrived Yet
The session’s closing discussion returned to a problem that has followed the AI infrastructure boom from its beginning: how do developers construct facilities today when GPU platforms and rack architectures are changing so rapidly?
Kestler challenged the assumption that each technology generation necessarily makes existing infrastructure obsolete.
Developers still need a durable basis of design. They still need to order generators, switchgear and substation equipment years in advance. And they still need to make investment decisions before knowing exactly which hardware will occupy a building.
The answer, Kestler argued, is not attempting to predict every future GPU. It is building enough adaptability into the underlying infrastructure to accommodate change.
“What will become obsolete is the inability for you to convert what you already have because you didn’t think through it first,” he said.
That may also be one of the clearest definitions of the execution challenge confronting the data center industry in 2026.
Demand remains immense. The technology will continue moving. New markets will continue emerging.
But the differentiator now is increasingly what can actually be financed, powered, entitled, staffed, built, adapted and accepted by the communities around it.
The projections brought the industry to this point. Execution will determine what gets built next.
