What the Ads Aren’t Telling You About Data Centers

The Hydrological Forecast is from the ICPRB, the water use data is from PW Water

A rebuttal to the industry’s polished pitch in Northern Virginia

If you’ve turned on a television in Northern Virginia lately, you’ve probably seen the glossy ads from Virginia Connects and other industry-backed campaigns. The message is soothing and simple: data centers are quiet neighbors, modest users of local resources, and generous contributors to schools, roads, and jobs.

That message is also incomplete. The ads ask residents to imagine “the cloud” as weightless, green, and nearly invisible. But here in Data Center Alley, the cloud has a physical footprint: transmission lines, substations, backup diesel generators, water withdrawals, stormwater runoff, and rising pressure on household utility bills. This is a rebuttal to the sales pitch and a reminder that Virginians deserve the whole story before being asked to subsidize the next wave of growth.

Claim 1: “Data centers barely use water.”

One of the industry’s favorite talking points is that many data centers do not use water for cooling “96% of the time.” That sounds reassuring until you look at what the number leaves out. It measures duration, not impact. In other words, it focuses on how many hours a facility may avoid evaporative cooling, while downplaying what happens during the hottest, driest, highest-demand hours of the year.

In Virginia’s climate, outside air may handle cooling much of the year. (It was real cold last winter.) But when temperatures climb in the summer and demand peaks, evaporative systems can consume large volumes of water precisely when rivers, reservoirs, and residents are under the greatest stress. The public deserves to know not just whether a facility uses water, but how much it uses during peak summer conditions, and that number is over 10% and growing.

The ads also tend to focus narrowly on onsite cooling. That leaves out the water consumed indirectly through electricity generation which is reported to be 90% of a data center’s water footprint. A data center that draws enormous power from the grid shifts part of its water footprint to power plants that cool turbines and equipment. Calling that “water free” is not transparency- it is accounting by omission.

Claim 2: “Water-positive pledges solve the problem.”

Major tech companies often point to “water positive” goals as proof that they are part of the solution. But water is not like carbon dioxide. A gallon saved in another watershed does not refill the Potomac, the Occoquan, or a local reservoir during a Virginia heatwave. The Interstate Commission on the Potomac River Basin (ICPRB) projects that by 2040 summer water use by data centers already in the pipeline with increase demand by 80 million gallons of water a day.

That geographic mismatch matters. If a facility’s local operations stress a local water supply, then the mitigation must be local, measurable, enforceable, and timed to the same season of demand. Otherwise, “water positive” becomes a public-relations phrase rather than a real safeguard for Virginia communities.

The same problem applies to offset-style projects that sound impressive in an advertisement but do little to address the specific burden placed on local residents. Virginians should ask a simple question: does the promised benefit happen here, when we need it, and at a scale that matches the impact?

Claim 3: “Dry cooling makes the resource problem disappear.”

Dry cooling can reduce direct water use, but it does not make the problem vanish. It often shifts the burden from water to electricity. Mechanical air-cooled systems rely on large fans and chillers, and those systems have to work hardest during the same heat waves when families are running air conditioners and the grid is under maximum strain.

So, when an ad suggests that a facility is environmentally benign because it uses less water, residents should ask the follow-up question: how much additional electricity does that choice require, and who pays for the generation, transmission, and reliability upgrades needed to support it?

Claim 4: “Efficiency gains will offset AI growth.”

The industry’s older efficiency story was built around conventional cloud computing. The AI boom changes the equation. High-density AI racks can draw far more power than traditional server racks, and they generate heat that is harder to manage with ordinary air cooling.

That means yesterday’s efficiency talking points cannot be used to wave away tomorrow’s load growth. If AI facilities require dramatically more power per square foot, the public should see updated, project-specific projections—not broad assurances built on the technology mix of the past.

Claim 5: “Backup generators are only for emergencies.”

Data centers are often promoted as clean, high-tech infrastructure, but many rely on large fleets of diesel backup generators. The industry describes these generators as rarely used, but their role has becomes more complicated. When the grid is strained and large customers are asked to reduce load or island themselves from the system. The grid is using them as additional power in their reserve calculations.

That raises a basic public-health question: if diesel engines are part of the reliability plan for a region packed with data centers, how are emissions monitored, limited, and disclosed to nearby communities? Residents should not have to accept unexamined pollution risks as the hidden cost of keeping the cloud online.

Demand-response payments and emergency operations also deserve scrutiny. If large facilities are compensated for reducing demand during crises caused in part by rapid load growth they brought, then regulators should ensure that ratepayers are not paying twice: once for the infrastructure buildout and again for temporary relief when that infrastructure proves insufficient.

Claim 6: “Data centers pay their own way.”

The ads emphasize tax revenue, but the cost side of the ledger is just as important. New substations, transmission lines, generation capacity, water infrastructure, road improvements, and emergency services are not free. When growth is fast enough, the question is not whether data centers pay something. The question is whether they pay the full marginal cost of the new infrastructure they require.

·         The ad version: Data centers generate tax revenue and support public services.

·         The missing context: Rapid load growth can require expensive new infrastructure, and those costs can place risk on households and small businesses unless regulators create strong protections.

That is why the debate over high-load rate classes and cost allocation matters. These policies are not anti-technology; they are basic consumer protection. If a facility needs enormous new grid capacity, ordinary residents should not be left carrying the financial risk if demand forecasts change or speculative projects fail to materialize.

Don’t Let Advertising Substitute for Accountability

The data center industry is not wrong that digital infrastructure matters. It does. But importance is not a blank check. Northern Virginia can support technology without accepting secrecy, weak siting rules, vague offsets, hidden water impacts, or utility bills that rise to finance private growth.

So, the next time a polished TV ad tells you that data centers are quiet, clean, and cost-free neighbors, remember what the ad leaves out. Ask for local water data. Ask who pays for the next transmission line. Ask how diesel emissions are monitored. Ask whether “water positive” means anything in our watershed. Ask whether the companies profiting from AI expansion are paying the full cost of the infrastructure they require. Virginia does not need more slogans. It needs transparency, enforceable standards, and a growth policy that protects residents first.