Some local legislators and residents in Northern Virginia, home to ~275 data centers with dozens more planned, complain of land use, cooling fan noise, and more
I have, my house is *in* that picture. — It's ridiculously bad. 24/7 constant droning even inside. — There's another one by Manassas Airport that doesn't have this issue. … Tweets: Paul Kedrosky / @pkedrosky : Northern Virginia is home to about 275 data centers, handling at least a third of the world's online use, with dozens more planned—and many locals are unhappy about it, especially the fan noise https://www.washingtonpost.com/ ... https://twitter.com/... Rona Kobell / @rkobell : Great story from @aolivo.Also an #environmentaljustice issue, and a land-use issue as these huge data centers create stormwater runoff and are built on farmland that could be open space for rec and flood-control. They also use huge energy users https://www.washingtonpost.com/ ... Fredrick Kunkle WaPo / @kunklefredrick : How development has turned parts of NoVa into a hellhole: Silicon Valley Chapter 2 ... deep dive by @aolivo into impact of cloud computing and data centers on people's lives https://www.washingtonpost.com/ ... Teo Armus / @teoarmus : The growth in NoVA's data center industry is sparking local land use debates in neighborhoods where data center buildings — some the size of several football fields — sit less than 100 feet from the nearest home.c By @aolivo https://www.washingtonpost.com/ ...
Context & Ripple Effects
When this story ran, Northern Virginia already held ~275 data centers carrying at least a third of the world's online traffic — the densest cluster anywhere — and the grievances on record (24/7 cooling-fan drone audible inside homes, farmland loss, stormwater runoff) looked like a local zoning quarrel rather than an industry-wide problem. Experts had flagged the underlying tension months later: with 2,701 US data centers in 2022 by Statista's count, finding the land and energy for new ones was already the binding constraint.
The intervening coverage turned this into a national template. Opposition has since grown sharp enough that 17 projects worth $98B were blocked or delayed in a single quarter, states like Arizona and Georgia have legislated restraints, and Loudoun County's board voted to slow construction despite the sector's outsized tax contributions. Even the specific complaint here — constant low-frequency noise — resurfaced years later as [[a:1171138|residents near AI-era campuses reporting infrasonic vibrations damaging their health and homes]].
First-order effects
- Local legislators in Northern Virginia gain a concrete, measurable grievance — round-the-clock fan noise and consumed farmland — that converts the dozens of planned facilities into contested rezoning and proximity-to-homes fights rather than routine approvals.
- Operators expanding in the world's largest data center cluster can no longer site purely on power, fiber, and land availability; residential adjacency becomes a live permit risk on every new proposal.
Second-order effects
- Other host counties confront the same tradeoff this cluster exposed: Loudoun County, with 250+ facilities generating large tax revenues, still voted under resident pressure to slow construction, setting a precedent that tax base does not buy automatic approval.
- The backlash generalizes geographically — towns across the US begin raising water and electricity objections to proposals, and statehouses including Arizona's and Georgia's respond with laws restraining the industry.
Third-order effects
- If siting friction keeps compounding, compute capacity increasingly migrates toward jurisdictions offering permissive zoning and cheap power, decoupling data center location from the demand centers they serve and reshaping which regions capture the tax and jobs.
- Environmental-justice framing of impacts — stormwater runoff, farmland loss, energy draw, noise — moves from complaint letters into formal permitting criteria, raising the cost floor for every future campus regardless of operator.
The trend: Data center siting is shifting from a technical power-and-fiber decision to a political one, with local social license emerging as a binding constraint on AI-era capacity growth.