When a Data Center Comes to Town

What changes for the people who live there? Compare the benefits, resource demands and real risks beyond the fence.

A data center can seem far removed from small-town life. Behind its walls, computers store and process the information that makes email, websites and other digital services work. Outside, people still need dependable water, affordable electricity, good jobs and a quiet place to come home.

Those worlds meet whenever a community considers a new campus. A construction contract can mean work for a family. Tax payments can support public services. A cooling system or a new electrical connection can change what the town must supply. A nearby household may experience the project quite differently from a business across town.

That is why the most useful question is specific: what does this project bring, what does it require, and how does it compare with other businesses the town could attract? Examples from Georgia and other states help answer it. They also show why one town’s experience cannot settle another town’s decision.

The first boom happens while people build

Construction and operation are different chapters. Virginia’s legislative research commission found that much of the industry’s economic benefit came during construction. Industry representatives described about 1,500 workers at peak construction of an individual building, compared with roughly 50 full-time workers—including contractors—for a typical operating facility of 250,000 square feet. Those are illustrative Virginia estimates, not staffing counts for every campus. [1]

Temporary work can still matter greatly. The question for a town is how much reaches its residents and businesses, what skills workers gain, and where they go when the building is finished. Permanent jobs deserve a separate count. Neither a statewide employment estimate nor a construction headcount tells us how many local families will earn a continuing paycheck.

There are tangible local investments beyond the construction site. In 2022, Mercer announced a $42,000 Google grant for workforce labs at its Douglas County academic center, with business training and digital-literacy activities among the planned uses. The grant is documented; participation and long-term outcomes require a follow-up. [2]

Follow the water all the way through

Douglas County offers a surprising connection: treated community wastewater helps cool Google’s computers. In its 2012 explanation, Google described diverting treated wastewater through an additional treatment system and into its cooling operation. Some water evaporates. The remainder receives further treatment before returning to the Chattahoochee. [3]

Using reclaimed water can reduce reliance on drinking water. It does not make evaporation disappear. That distinction matters when a community asks both about its drinking-water supply and the water returning to a river.

Cooling design matters too. Virginia’s review found wide differences in facility water demand. A single industry-wide gallon figure cannot describe them all. Ask where the water comes from, how much is used during hot weather, how much returns, and whether the number describes construction, expected operation or an actual meter reading. [1]

The tax payment matters more than the investment headline

A large investment can create a valuable tax base. How much becomes public revenue depends on what is taxable, applicable rates and any incentive agreement.

Loudoun County, Virginia, says data-center revenue has helped support schools, emergency services and other facilities while enabling tax relief. The county also warns against becoming overly dependent on that revenue. Its experience shows both the opportunity and the responsibility of building recurring public spending around a concentrated industry. [4]

For any town, the useful comparison is between taxes and other payments actually received, taxes forgone under agreements, and public costs associated with the project. A press release announcing billions in investment does not supply that comparison. Nor does an abatement figure, by itself, establish what would have happened without the incentive.

Who pays when the power system grows?

In West Georgia, GreyStone Power describes requiring data-center customers to pay upfront for dedicated infrastructure, meet minimum billing requirements and provide financial security. These measures are intended to protect existing members when large new customers need substantial power. [5]

That is an important benefit of careful contracting. It is also a reason to read the terms. A utility’s general policy does not tell readers what every individual project has agreed to, or establish the eventual effect on a household bill.

The practical questions remain useful everywhere: who pays for added infrastructure, what payment obligations begin before full operation, and what happens if the expected customer demand never arrives?

Compared with what?

A warehouse, a factory and a data center can occupy similar-looking buildings while placing very different demands on a town. The evidence supports several distinctions. It does not yet support a single score declaring one the best use of every site.

Water and employment also need their own comparisons. A cooling design can change water demand dramatically; a large building need not employ a large permanent workforce. We should compare the actual proposed operations and payroll, not assume all warehouses or all factories are alike.

The neighbor’s experience belongs in the story

Virginia’s review found that some residents experienced persistent noise even where it rarely exceeded ordinary noise limits. It also found that most facilities did not generate noise complaints. Site selection and design matter. [1]

For a nearby household, the relevant evidence is what happens at that property: measurements, timing, documented complaints and the response. A countywide economic benefit does not answer a particular noise complaint. A complaint at one facility does not prove every facility creates the same problem.

Before construction, residents can ask for a baseline sound study and clear operating limits. After opening, they can ask how conditions are checked and who handles a problem. The same attention should extend to lighting, drainage and the view from neighboring homes.

Plan for a change of plans

Computing operations can move. A contractor case study describes consolidating Cigna/eviCore workloads from Atlanta and Dallas into other locations and decommissioning the old data centers. It does not establish what happened to the Atlanta building or local employees. [6]

A community therefore needs to understand more than the opening-day promise. Who remains responsible if the owner changes? Can the building be reused? What happens to unfinished work or infrastructure built for demand that never appears?

What the evidence supports

Our assessment: these examples do not establish that data centers are uniquely harmful—or that they are the best choice for every town. They do establish that energy demand deserves particular scrutiny. Their public-revenue benefit can be substantial. Water demand and neighborhood effects depend on the design and location, while alternative businesses bring their own resource and pollution burdens.

A town has a stronger case for welcoming a project when credible public payments survive the incentives, sufficient water and power are available, the customer bears its attributable infrastructure costs, and neighborhood protections can be enforced. A project that cannot demonstrate those conditions has not yet earned the same confidence.

This is a conditional judgment, not a finding that every data center is beneficial or that a particular West Georgia campus has passed all those tests. We still lack matched local figures needed to declare a warehouse, factory or data center the overall winner. The evidence supports judging individual proposals by their documented benefits, costs and protections.