How Big a Price Cut Does It Really Take to Sell a House Next to a Data Center?

The 40% guess versus what the studies actually found

So I went looking for the real haircut, expecting to confirm the doom. Here's the first thing that stopped me: the biggest recent study, out of George Mason's regional analysis center in 2025, looked at thousands of home sales across Northern Virginia's data-center belt. After they controlled for a dozen other things that move a price, homes closer to data centers actually sold for a little more, not less. Not a typo. More.

Realtor.com ran its own version and got the calmer cousin of that finding — in the two years after a data center switched on, nearby ZIP codes basically tracked the towns next door. No real gain, no real loss. On paper, the 40% haircut doesn't exist. On paper, there's barely a haircut at all.

Which is where I almost closed the tab and wrote "good news, everybody." But that word — average — kept nagging at me. An average is what you get when you throw the guy whose house went up in with the guy who can't get a single showing, and let them cancel each other out. The average feels fine. Ask the second guy how he's doing.

Why some people swear their house is now worthless

The gap is distance, and it's brutal how fast it kicks in. A mile away, a data center is a rumor — a big beige box you drive past. Under a quarter mile, it's a neighbor that runs cooling fans every hour of every day, and the market notices. Analysts describe it plainly: properties right up against an operating facility take a real valuation hit from the noise, the truck traffic, and the sheer industrial look of the thing out your kitchen window.

I found one Northern Virginia listing that made it concrete. A four-bedroom sitting about two football fields off a data center fence line went up for roughly a hundred grand under the near-identical houses a few streets over. That's not 40%. But it's not nothing either — it's a car, a kid's first two years of college, gone because of what got built across the road after you signed your mortgage.

So the commenter wasn't crazy, just aimed at the wrong house. The 40% panic and the "no impact" studies are both true. One's describing the neighborhood. The other's describing the one poor place with a view of the loading dock.

The number nobody quotes: the house that just doesn't sell

Here's the part that actually rattled me, because a discount at least assumes a buyer. What kept coming up in the noise complaints wasn't a lower price — it was no price. Homeowners near these sites describing the sound like tinnitus you can't turn off, or louder than standing at an airport. One neighbor called it a refrigerator the size of a Walmart that never once cycles off. You can drop your price all you want; some buyers hear that on the walkthrough and they're just done.

And it's spilling into court. A homeowner in Louisa County, Virginia is suing Amazon over a data center going up next door — cracked foundation from the vibration, tap water that ran muddy brown, and, in his own words, a house he can't sell. Neighbors in Michigan and New Jersey have filed over the hum. Chesterfield County paused a project mid-fight. When your escape hatch is a lawsuit, "what discount do I offer" isn't really the question anymore.

That's the thing the tidy percentage hides. The cost of the truly-adjacent house isn't always a smaller check at closing. Sometimes it's the year of your life spent sitting on a listing nobody will touch.

So what would I actually demand — and where I got stuck

If I'm the buyer, I stopped caring about a flat percentage the second I understood the map. I'd want to know exactly how close, which direction the fans point, whether it's running yet or still under construction — because the dust-and-hammering phase is its own separate misery — and what the county already has on the docket to build next. A house a mile out with good tree cover is a normal house. A house on the fence line is a coin flip on whether it sells at all.

The one thing I couldn't pin down, and I looked, is a clean nationwide "here's your discount" number for the fence-line houses, because there just aren't that many sales yet and every site is its own animal. What I could pin down is that the honest range runs from "basically zero" to "can't sell," and which end you land on is mostly about how far your bedroom is from the loading dock.

This is the exact rabbit hole Byte Bungalow keeps falling down — the power, the water, the hum, and who ends up paying for it all. If you want the fuller version of this one, come watch the video that goes with it and subscribe on the channel. That's where the next thread gets pulled.

Common questions

Keep reading

Sources

• cra.gmu.edu — George Mason Center for Regional Analysis, "Data Centers and 2023 Home Sales in Northern Virginia" (Aug 2025): after controlling for 13 other factors, proximity to data centers cor
• ownluxuryhomes.com — NoVa case study: 4-bed ~200 ft from a data center listed at $580,000 vs comparable homes $685,000–$710,000 = ~15–18% discount; buyer concern = "ever-present whirring hum" from 24/7
• stocktitan.net — Realtor.com report: share of home sales within 5 mi of a ≥50MW data center rose from 0.67% (2018) to ~1.5% (2026), projected 2.3% by 2027; in two years after activation, home value
• cbsnews.com — Chicago South Loop residents fighting a data center; cooling fans run 24/7; neighbors say they can't sleep.
• housingwire.com — (attempted fetch, HTTP 403; used only headline framing: data centers as real estate's "newest pricing wildcard").