Speed to Compute: Outreach Strategy Now Determines the Data Center Schedule
-
August 25, 2026
-
On the fastest data center projects, protecting the timeline begins long before breaking ground. The effort starts in the community, where site selection, stakeholder outreach, and non-technical risk assessments happen before a location is finalized. A project can clear every technical hurdle — power secured, cooling system designed, construction plans approved — and still miss its launch date because a historic site went unsurveyed, a noise complaint unanticipated, or a community never consulted.
Most of the data center capacity under construction today, 77 percent, already has a committed user, leased to a tenant or built by its future occupant. Yet the Financial Times reports that roughly 40 percent of projects due to open in 2026 are at risk of slipping at least three months behind schedule, and 2027 does not look better. When a customer commits months in advance, a delay isn’t an inconvenience; it is a broken promise with financial and public consequences.
The threats to speed have moved off the engineer’s renderings and into the neighborhood. From January to March 2026, 75 data center projects worth $130 billion were blocked or delayed, and local opposition groups more than doubled, from 396 to 833 across the country, according to Data Center Watch. Policymakers introduced more than 300 bills addressing data centers in over 30 states in the first six weeks of 2026 alone. Developers who treat community and regulatory engagement as an afterthought start behind.
The Playbook Needs Updating
A year ago, most developers didn’t study the community until after buying the land. Today, the most responsible developers start their efforts in the neighborhood, even before buying a parcel, mapping the local landscape: who will support or oppose the project, which issues carry weight, and what complications exist nearby. Engage early, and the site or plans can adapt. Engage late, and the schedule could fall behind.
This summer in Virginia, a county board halted a 99-acre data center after a two-century-old cemetery was found on the property — spotted, by the developer’s own account, only after the land was purchased some 20 years earlier. Surfaced during site selection, this points a developer to a different parcel. Surfaced in the approval process, it becomes months of delay.
A state with an easier regulatory environment doesn’t help. Political risk follows the project, and the newer markets are already producing their own opposition groups and moratorium bills.
This is where the discipline of capital projects earns its keep. Engineers know the influence curve: Changes are cheap before the design is fixed and expensive after. Well-run projects collate every known risk at the start in a single register — each with an owner, a mitigation, and a retirement date. What needs updating isn’t the discipline but the register’s contents: The cemetery survey, the noise-sensitive neighbor, the water narrative, the permitting pathway, and the local labor picture belong alongside geotechnical findings and long-lead equipment. Collated up front, they’re line items. Discovered late, they are the schedule.
Community Input Belongs in the Design
Most developers treat community feedback as a messaging problem: hear the concern, craft a response, hold the line. The best treat it as an engineering input, changing what gets built, not just what gets said about it.
Take noise as an example. A developer can position generators away from neighbors, add enclosures, stay within the legal decibel limit, and still lose. The fight is not usually about decibels; it’s the low hum the ordinance never measures, but neighbors cannot tune out.
The same holds for power. Many campuses turn to on-site generation to skip grid-connection waits that can run up to seven years in the busiest markets, and the air-emissions permits that generation requires are tightening fast. Handled late, permitting stalls everything: a 12-to-18 month delay on a billion-dollar build burns $70 to $105 million before earning one dollar. Handled early, it simply shapes the design — turbine selection, placement, and emission controls.
Address non-technical risk only after the design is set, and a developer can merely defend a fixed plan. Address concerns during siting, and developers keep options: relocate generators, adjust setbacks, change equipment, before pushback becomes a costly, public fight.
Clearing a permit is not the same as earning acceptance. The rules outline what a developer can build, but not what the people next door will organize against.
Being Right Is Not Enough
A developer can have every fact right and still lose the room. Water is where this happens most, and the trap is partly self-made.
Permitting forces a project to request water for its maximum, fully built-out capacity, because public water systems must guarantee peak demand. This ceiling amount enters the public record, and the community reads it as the everyday number. In one town with under 2 million gallons of spare capacity, a single facility’s peak request of 1.2 million gallons a day reportedly triggered a water infrastructure upgrade exceeding $100 million. Numbers like that harden into the story before the developer says a word.
Everyday reality is usually far lower. Modern closed-loop cooling, now standard for dense artificial intelligence (AI) hardware, recirculates the same water in a sealed system, like a car radiator, rather than evaporating it away, cutting daily freshwater use by as much as 70 percent versus older methods. Many modern data centers also can rely on outside air for cooling, and not need water most of the year. But once the maximum need is on paper, explaining that real use is a fraction of it sounds like a company walking back its own numbers, defending a figure the process required it to publish, to a community that already read it as a threat.
The water question is not new; it is a centuries-old industrial trade-off applied to modern technology. One Virginia paper mill draws about 22 million gallons of groundwater a day, and even after a 10-year cutback will still use roughly 17 million. That is many times the one-to-two million gallons a day a large data center’s peak permit typically requests, the figure that alarms communities.
Developers tend to skip one more chapter: for a modern closed-loop campus, the build is often the thirstiest phase, drawing hard on local supplies for a season — dust control, soil conditioning, concrete, and the one-time fill of the cooling loops. South of Atlanta, on one of the country’s largest campuses, crews drew tens of millions of gallons before it came to light, after neighbors noticed a drop in water pressure. Construction water is temporary and has an end date, but a community that discovers it by surprise does not experience it that way. Volunteer the story in two chapters: a heavy, temporary construction draw with an end date, then a modest operating draw for the asset’s life. Disclose the thirsty chapter and earn the right to be believed on the quiet one.
Electricity is the hardest version of this story, because the public alarm isn’t baseless — bills are rising. But the causes are mostly structural and predate the AI build-out: aging grid infrastructure, costly transmission and distribution upgrades, weather hardening, and volatile fuel prices. Data-center demand adds to that pressure, and in a few high-concentration markets it is a real factor — but it is one driver among many, and rarely the largest. Arguing that data centers bear no blame at all collides with people’s actual bills and loses credibility instantly. The credible move: separate the structural cost drivers from the project’s own footprint, then put the project’s picture in the open early — what it will draw, what it will fund, and what it will not shift onto residents — before the feared number becomes the only number.
Accurate numbers don’t defend themselves. Presented for the first time at a public hearing, they aren’t informing the community — they’re rebutting claims, and rebuttals often lose.
The Fastest Build Can Be the Slowest Project
A developer’s instinct is often to outrun opposition: tighten the stage-gate process, modularize, build components off-site, and move faster than anyone can object. Modular construction does compress timelines — by roughly 30 percent, per one industry estimate, largely by shifting assembly into a factory. But speed bought wrong can cost more than it saves. In contested markets, the local building trades, including electricians, pipefitters, and laborers, are often a project’s most credible advocates, the voices to whom a county board listens. Manufacturing capacity elsewhere and trucking it in can remove that constituency, hardening the opposition it was meant to avoid.
The answer is not to abandon modular construction but to pair it with a real local stake. Off-site manufacturing still needs local workers, and a proactive workforce program, training and hiring local building trades before a project needs them, turns labor into an ally. Done well, the community gains jobs, a more stable workforce, and a broader tax base. Those are not community-relations niceties, but scheduling assets that make the next project easier to site.
How Stage Gates Speed the Build: “Slow Is Smooth, Smooth Is Fast”
Most developers say they run a stage gate process: Capital doesn’t advance until the project proves ready, with gates at site selection, concept design, final investment decision, and construction release. In practice, this machinery lives mostly on paper. Teams work in silos, information reaches the gate in fragments, and no one collates the risks in one place; as a result, items are discovered too late. Nowhere is this costlier than the investment decision, where capital expects a complete picture, gets a partial one, and sends it back — a gate meant to be passed once becomes a cycle of resubmission that burns the calendar before a shovel moves. And the gates typically measure geotechnical work, power, and budget. Community and regulatory readiness rarely appear, so a project can clear every gate and still not be ready in the arena that now sets the schedule.
The fix is twofold. First, run the gates as designed: one integrated decision package per gate, every risk in a single register before the meeting, one decision made once. Second, give non-technical risk the same standing as technical risk — no investment decision without a permitting pathway, an engagement record, and the resource story already in the open and no construction release without the local workforce plan. Gates keep a project from outrunning its own information. Broaden what they demand, and the risks collated at the start are retired on schedule, instead of surfacing at a town hall.
Conclusion
The biggest risks to data center schedules now lie in the communities where projects are built. A cemetery, a noise profile, a water permit, an electric bill, or a local labor pool can all be addressed early — or cause costly delays when discovered too late.
The developers who consistently hit their targets have adjusted: they no longer treat community and regulatory engagement as work that follows engineering, but as part of it: gated, owned, and scheduled before a site is ever selected.
Community engagement is not a brake on speed to compute. It is what protects it.
Related Insights
Published
August 25, 2026
Key Contacts
Senior Managing Director, Global Head of Data Center Construction, Projects & Assets
Senior Managing Director
Managing Director Americas Head of Telecom, Media & Technology Sector
Managing Director