04 / TECH WITHOUT EMPATHY
Why Towns Are Saying No to Data Centers: From 85% to 11%
The projects had capital, land and power plans. What they didn't have was the town's trust.
DATA CENTER APPROVALS BY US CITY COUNCILS, 2022 VS. 2025–26
In 2022, when a data center proposal reached a city council, the answer was almost always yes. Eighty-five percent of the time.
By 2025 and 2026, the same vote came back yes just 11% of the time.
For the book, my team mapped 93 of those council battles across 24 states. We kept tracking. The AsimovGrid tracker now counts more than 135 cases of community opposition in 38 states: denials, moratoriums, lawsuits and new state laws.
The engineering didn't get worse. The money didn't dry up. What changed was the room. Residents stopped believing the deal was being made for them.
The pattern everyone recognizes
You've seen this movie. A tech company announces a new data center. The governor holds a press conference. There are promises of hundreds of jobs, maybe thousands. Tax incentives are offered. Land is rezoned. Environmental reviews are accelerated. The ribbon is cut.
Then, quietly, the local grid starts to strain. Electricity prices for residents tick up. Water restrictions arrive during droughts, but the data center keeps humming. The roads around the site wear down under construction traffic. And the jobs? Mostly temporary construction work. The finished facility might employ 50 people.
Residents have now watched that movie enough times to know the ending. So they've started showing up before the opening scene.
What the denials actually say
When we read through the council records, the reasons for saying no were remarkably consistent, and almost none of them were about the technology.
Noise and environmental impact: diesel generators, industrial cooling systems and the constant low-frequency hum. The character of the town: residents who wanted to preserve rural landscapes, heritage and identity. And infrastructure strain: power, water and the fear of higher utility bills.
That last one is the reasonable fear. A single hyperscale data center uses as much electricity as 80,000 homes. Data centers are projected to go from roughly 4% to about 12% of US electricity by 2030. A single facility can use between 1 and 5 million gallons of water a day for cooling. In The Dalles, Oregon, Google's data centers used more than a quarter of the city's water supply in a single year. In Loudoun County, Virginia, home to the largest concentration of data centers on the planet, residents are fighting the hum of cooling systems and the conversion of their rural landscape into an industrial corridor.
Now put those numbers next to cities that can't keep clean water flowing or fix a water plant that's nearly a hundred years old. We are adding the bones of a new digital economy on top of a skeleton that is already brittle. That's the maintenance trap on steroids.
The brain is eating the bones. And nobody seems to notice until the lights go out.
This is a trust problem, not a PR problem
It's tempting to read 85% to 11% as a messaging failure. Better renderings, a bigger community fund, a sharper slide at the hearing. That reading is wrong.
Trust in government is at historic lows. The 2025 Edelman Trust Barometer found that only 42% of Americans trust the federal government to do what is right, and 61% of people globally feel a moderate or high sense of grievance, a belief that their government makes their lives harder and serves only the wealthy. City council meetings that used to be sleepy have turned into theater, and data center votes sit right alongside every other hot-button fight.
When residents don't trust city hall, they don't trust the deal city hall negotiated. The tax break, the utility rate, the job promise: every term gets read as a favor to someone else. No amount of marketing closes that gap after the fact. If you're on the investment side, I wrote about what this means for deal risk in $170 Billion Stalled.
What would change the vote
The towns saying no aren't asking for the impossible. They're asking for straight answers, early.
- Answer the six questions before anyone asks. How much water will the facility use? How much power? From what source? What tax breaks is it getting? What are the terms with the local utility? How many jobs will go to local residents? These are public records anyway. Publish them first.
- Be honest about jobs. If the finished facility employs around 50 people, say 50. The gap between the press-conference number and the real one is where trust dies.
- Pay your share of the bones. Residents assume a data center will raise their bills and drain their water. Show, in the deal itself, how the project contributes to the grid and water systems everyone else depends on.
- Bring your own power where you can. On-site generation and microgrids take pressure off a strained grid and remove the single biggest fear in the room.
- Listen in public, and answer in writing. The book's Law 21, The Loud Civic Voice, asks cities to document a response to every piece of resident feedback. Developers who do the same, before the hearing, stop being the outsider in the room.
Residents who push back on data centers are not anti-technology. They are pro city.
And data centers are only the newest test. The same soft systems (trust, governance, honest information) decide whether a city can replace a century-old water plant or keep insurers from leaving. That's the argument of the whole book, and it's why infrastructure fails after the money is raised.