The new poll found that 41 percent of Americans would oppose a data center being built within three miles of where they live, up from 28 percent in the January survey. The number supporting those projects dropped during the same period, from 37 percent to 24 percent.
The shift was especially pronounced among people who said they had voted for Kamala Harris in the 2024 presidential race.
The results reflect the scope of the challenge facing the industry and its supporters as the rapid buildout of data centers brings the issue home to voters ahead of the November midterms. More than 200 data centers are going up in dozens of competitive House districts, a recent POLITICO analysis found.
“It’s rare to see political attitudes change this fast in an era where they are often quite static and partisan,” said Seb Wride, head of polling at Public First.
Not surprising at all. There is paid for the locals when these things appear, but they don’t get anything back for that paid. Even if they’re computer literate and using ai, there’s no notice that this came from their local centre. Nor is there anything that’s unique to a particular location.
People get annoyed at being inconvenienced so that others, elsewhere can make money.
Here in Michigan there are large industrial facilities that most locals have no idea what is done inside. There's noise, electric usage, trucks coming in and out, but most people think it's ok. Data centers are seen differently.
It’s different because those places typically provide jobs for the people in the surrounding area. Data centers don’t employ enough people to provide an economic benefit for the people that live near them.
If you live in the area, it's easy to say that. There's no protests in the local parks. There's no shouting matches at city council meetings.
I say this as a person living in a town where they are building a bunch of data centers, and nobody cares about the existing industrial areas which include factories and heavy industry. No protests, no shouting matches at city council meetings. None of that. Granted, those were built a long time ago.
Most people don't show up to the data center protests. Most people probably are probably not aware of planned industrial development in the rural north part of my town. Most people probably don't read enough local news to be aware. My general feeling. I haven't done a count, but gut feeling maybe 1% of homes and businesses in my town have "Say NO to Data Centers" signs.
The tech industry has spent that last few decades fucking with people: addictive infinite scrolling sites, dark patterns, etc.
Now the same assholes are saying their machines will eradicate the job market. And also will make their neighborhood noisy and their water and electric bills high.
The industry has earned this pushback several times over.
(I am generalizing but you get the idea from a layperson's perspective)
States may dangle the tax carrots, but localities could be the ones paying for road and utility improvements. The costs for that could be much less than projected gains in tax revenue, but that's a gamble. Not to mention being left with abandoned buildings. We didn't even need to have a bubble pop to see abandoned facilities. If one company goes out of business (normal in a healthy economy), there's little we can (and politicians are willing to) do to make sure defunct companies clean up any messes.
From the few cases I have looked at, tax revenues for the towns is in the $30-40M/yr range.
I cannot fathom what kind of road improvements a datacenter would need that wouldn't be easily offset. Or what demolition costs for a square box would be that would eclipse that, nevermind the value of the underlying land.
This could have been stopped so easily on its tracks with a simple premise: 1 day turnaround build permits, to any proposed off-the-grid data centers.
BYO power. BYO water. BYO pipes & sewage. OK, you are good. Now build!
Instead they are look like parasites invading the heartland and sucking resources (true or not), and owners pretend to be surprised when they are getting voted down.
> 1 day turnaround build permits, to any proposed off-the-grid data centers.
Utility planning is one tiny part of approving a building permit. You still need to look at storm water runoff, fire safety, environmental impact etc…
I’m also not even sure what bring your own water even looks like.
And even if they are providing their own wastewater treatment, that requires an even closer look than hooking up to municipal sewage. It also requires at least occasional long term inspection because these kinds of systems fail at insanely high rates.
Part of the problem is xAI. I bet the first reaction of many people when they hear someone wants to build a data center in their area is "Oh no...they want to do to us what they did to those people in Tennessee!" [1].
They built it in a poor, 90-95% Black, area that already had air quality problems leading to high rates of asthma and respiratory illness, and then decided they could not wait for the power company to upgrade infrastructure for them so they powered it with dozens of unpermitted gas turbines that lacked pollution controls, massively increasing the air quality problems and health problems there. (They are building their own gas power plant which will have permits and satisfy state environmental regulations but still will emit a lot of pollution which will continue to be a big problem for the residents).
Another part of the problem is that even projects that seem to be a good deal push so hard for fast approval it makes people suspicious.
Consider Project Jupiter, a massive (1400 acre mega data center) project in a poor area of New Mexico. From the NPR story I heard about this one it seems it might actually be a good deal for the area.
The public knew for months it was being planned but the details (400 pages of complex contract details and tax documents with many things still blank) were only released to the public 4 days before the County Board of Commissioners voted on it (4-1 in favor).
A data center on a ship out in the ocean is orders of magnitude cheaper and orders of magnitude more capable. If we can't demonstrate a solution to the heat problem, in-orbit DCs that contribute meaningful compute are a fantasy and nothing more.
You don't in fact have viable solutions to this problem when orbital launch prices are above like 100USD per ton, probably needs to be closer to 20USD per ton.
I choose to accept reality. Until someone can do it, it hasn’t been done. Once that happens, this will cease being vaporware. The bubble is filled with vapors and those who drank the koolaide can’t seem to get themselves to accept.
Solar power in space is not cheap. It's efficient. A solar panel in space produces 4-10X as much power as one on land.
Which would be relevant if solar panels or land were scarce and expensive. Neither are. It's easier and cheaper to put 4-10X the number of panels on land.
I've yet to hear of any solution to the heat problem, only vague handwaves that it will be solved. But it's the single largest and most challenging obstacle (amongst a set of large, challenging obstacles)
My searches aren't turning anything up, would you mind sharing what you know about solutions to the heat problem or point me towards a resource if can learn from?
It still doesn't make any sense: if they're not going to use bodies of water to cool them in space, they can not use a body of water to cool them in the middle of nowhere. The same with the solar panels
The water moves the heat from the DC to some near-infinite sink (ocean, river, atmosphere if evaporative, whatever.) In space it's quite (very) hard to move the heat anywhere. So water cooled ocean DC is more realistic than an orbital DC.
If they were talking about putting them on the moon I'd get it; use the moon itself as a giant heat sink. Latency isn't ideal but for applications where that's not a problem (like AI; 3s round trip isn't meaningful in the context of a chat session) then why not.
No they aren't because it was never about the externalities of the data center itself (for most - it was for a few people who actually were screwed by the noise, for example). Instead, it's because most of this crowd are anti-AI luddites who hate that change is happening too fast for their minds to comprehend or adapt to.
They stand in the way of a nearly limitless number of positive outcomes from this technology and directly have blood on their hands that they wipe away with thought terminating cliches like "precautionary principle"
Or look at it this way, a massive change has been forced on the public through the decisions of a small number of people. Basically the whole "Accept your fate, it is inevitable" backfired spectacularly. The people were robbed of their agency, their democratic process. This is just the democratic will reasserting its agency.
I think that ignores the scale and speed of what is happening with the buildout. That it's gotten to the point that states are putting in moratoriums, thats pretty incredible. I've never seen political government forces mobilize this quickly except in some sort of emergency situation.
It depends on the scale. If space data centers become a thing at the scale Musk and others are talking about (millions of space data centers) then rockets will indeed became a major issue for the atmosphere, as will the space data centers themselves.
The total atmospheric pollution from rockets, even at that scale, will be small compared the total from terrestrial sources but unfortunately atmospheric pollution works like real estate: location matters.
Orbital rockets deposit stuff into the upper stratosphere which is about the worst place you can put. Things there have effects massively disproportionate to the same amount of the same things emitted into the lower atmosphere by terrestrial sources. Some have up to 500x the effect when emitted in the upper stratosphere than when emitted by airplanes or on the ground.
It's been OK so far because in the entire history of rockets there have only been about 7500 orbital launches (which are what matter for the upper stratosphere--see below). With a million space data centers which are expected to each have a useful life of 5 years there would have to be about 2000 Starship launches (@100 data centers per launch) per year just for replacing the ones that have reached EOL. That's after the 10000 launches for initial deployment.
Worse, satellites damage the upper stratosphere on reentry. With a million space data centers that's 200k reentries per year. The cumulative number of reentries so far is around 10k satellites and spent rocket bodies, and around 25k if we count smaller things that we can track like debris.
There have been tens of thousands of suborbital launches that have reached the upper stratosphere but most of them have been sounding rockets which are small enough that the cumulative pollution from all of them is less than a year's worth of current orbital launches. On top of that most of the sounding rockets were for weather monitoring and have been replaced by balloons. Current sounding rocket use is negligible. There have also been ballistic missiles but by the time those are in the upper stratosphere they are ballistic and don't pollute much. Hence it is only orbital launches we have to worry about.
An 84yr old who doesn't understand the technology and 72yr old political hack? Both of whom are populist and lived in the fringe most of their careers.
Data centers would be less contentious in the U.S. if energy was plentiful and clean. As it is, it externalizes costs onto the public and strains profit-optimized electrical grids while contributing to local pollution and global warming.
The question isn't whether that's true, the question is the magnitude of the effect.
By all accounts, it's completely blown out of proportion. If people genuinely, actually care about energy costs, they can vocally support when calls for setting your thermostat at 78F in summertime go out.
The same demographic is against building new power plants, especially in their vicinity. There could have been abundant power if Nixon‘s plan to build out nuclear had been allowed to proceed.
The new poll found that 41 percent of Americans would oppose a data center being built within three miles of where they live, up from 28 percent in the January survey. The number supporting those projects dropped during the same period, from 37 percent to 24 percent.
The shift was especially pronounced among people who said they had voted for Kamala Harris in the 2024 presidential race.
The results reflect the scope of the challenge facing the industry and its supporters as the rapid buildout of data centers brings the issue home to voters ahead of the November midterms. More than 200 data centers are going up in dozens of competitive House districts, a recent POLITICO analysis found.
“It’s rare to see political attitudes change this fast in an era where they are often quite static and partisan,” said Seb Wride, head of polling at Public First.
Not surprising at all. There is paid for the locals when these things appear, but they don’t get anything back for that paid. Even if they’re computer literate and using ai, there’s no notice that this came from their local centre. Nor is there anything that’s unique to a particular location.
People get annoyed at being inconvenienced so that others, elsewhere can make money.
Here in Michigan there are large industrial facilities that most locals have no idea what is done inside. There's noise, electric usage, trucks coming in and out, but most people think it's ok. Data centers are seen differently.
It’s different because those places typically provide jobs for the people in the surrounding area. Data centers don’t employ enough people to provide an economic benefit for the people that live near them.
"but most people think it's ok."
How would you know what "most people" think? Broad generalizations and opinion stated as fact.
If you live in the area, it's easy to say that. There's no protests in the local parks. There's no shouting matches at city council meetings.
I say this as a person living in a town where they are building a bunch of data centers, and nobody cares about the existing industrial areas which include factories and heavy industry. No protests, no shouting matches at city council meetings. None of that. Granted, those were built a long time ago.
That's a few people, not most people.
Most people don't show up to the data center protests. Most people probably are probably not aware of planned industrial development in the rural north part of my town. Most people probably don't read enough local news to be aware. My general feeling. I haven't done a count, but gut feeling maybe 1% of homes and businesses in my town have "Say NO to Data Centers" signs.
The tech industry has spent that last few decades fucking with people: addictive infinite scrolling sites, dark patterns, etc.
Now the same assholes are saying their machines will eradicate the job market. And also will make their neighborhood noisy and their water and electric bills high.
The industry has earned this pushback several times over.
(I am generalizing but you get the idea from a layperson's perspective)
Usually the states give the incentives, the local towns usually get their tax revenue doubled or tripled, hence why these boards sprint to approval
States may dangle the tax carrots, but localities could be the ones paying for road and utility improvements. The costs for that could be much less than projected gains in tax revenue, but that's a gamble. Not to mention being left with abandoned buildings. We didn't even need to have a bubble pop to see abandoned facilities. If one company goes out of business (normal in a healthy economy), there's little we can (and politicians are willing to) do to make sure defunct companies clean up any messes.
From the few cases I have looked at, tax revenues for the towns is in the $30-40M/yr range.
I cannot fathom what kind of road improvements a datacenter would need that wouldn't be easily offset. Or what demolition costs for a square box would be that would eclipse that, nevermind the value of the underlying land.
>It’s rare to see political attitudes change this fast
On the contrary - these mass political crazes are like flash-floods these days.
This is not your standard partisan issue.
What the Unabomber, Steve Bannon’s tech guy, and Bernie Sanders taught me about the great data center backlash of 2026 - https://www.wired.com/story/how-data-centers-broke-american-...
Data center approvals in Texas halted until audits completed, Gov. Greg Abbott says - https://www.texastribune.org/2026/08/03/texas-data-center-pr...
DeSantis lambastes White House over possibility of blocking state-level AI laws - https://www.washingtonexaminer.com/policy/technology/4605403...
New York becomes the first state to impose a data center moratorium - https://www.reuters.com/world/new-york-becomes-first-state-i...
Libertarian Ohio governor candidate calls for ‘immediate moratorium’ on new data centers -https://www.cleveland.com/news/2026/08/ohio-governor-candida...
Oregonians want data center moratorium and no more tax breaks, poll finds - https://www.oregonlive.com/silicon-forest/2026/08/oregonians...
Virginia state senators call for data center moratorium - https://www.nbcwashington.com/news/local/northern-virginia/v...
Tucker Carlson, Kevin O’Leary have testy exchange over Utah data center issues - https://www.businessinsider.com/tucker-carlson-kevin-oleary-...
"Tucker Carlson, Kevin O’Leary"
A real meeting of the minds.
This could have been stopped so easily on its tracks with a simple premise: 1 day turnaround build permits, to any proposed off-the-grid data centers.
BYO power. BYO water. BYO pipes & sewage. OK, you are good. Now build!
Instead they are look like parasites invading the heartland and sucking resources (true or not), and owners pretend to be surprised when they are getting voted down.
> 1 day turnaround build permits, to any proposed off-the-grid data centers.
Utility planning is one tiny part of approving a building permit. You still need to look at storm water runoff, fire safety, environmental impact etc…
I’m also not even sure what bring your own water even looks like.
And even if they are providing their own wastewater treatment, that requires an even closer look than hooking up to municipal sewage. It also requires at least occasional long term inspection because these kinds of systems fail at insanely high rates.
> BYO power
Like 26 gas turbines that make a ton of noise and pollution?
> BYO water
What? How does one BYO water?
> BYO pipes & sewage
Pipes and sewage to where? There's an implied common ground at the end of those.
Yes
It falls from the sky
Septic tank
BYO power with a carbon and/or noise and/or local emissions footprint worse than the grid is definitely not good.
Part of the problem is xAI. I bet the first reaction of many people when they hear someone wants to build a data center in their area is "Oh no...they want to do to us what they did to those people in Tennessee!" [1].
They built it in a poor, 90-95% Black, area that already had air quality problems leading to high rates of asthma and respiratory illness, and then decided they could not wait for the power company to upgrade infrastructure for them so they powered it with dozens of unpermitted gas turbines that lacked pollution controls, massively increasing the air quality problems and health problems there. (They are building their own gas power plant which will have permits and satisfy state environmental regulations but still will emit a lot of pollution which will continue to be a big problem for the residents).
Another part of the problem is that even projects that seem to be a good deal push so hard for fast approval it makes people suspicious.
Consider Project Jupiter, a massive (1400 acre mega data center) project in a poor area of New Mexico. From the NPR story I heard about this one it seems it might actually be a good deal for the area.
The public knew for months it was being planned but the details (400 pages of complex contract details and tax documents with many things still blank) were only released to the public 4 days before the County Board of Commissioners voted on it (4-1 in favor).
The public was not happy with that timeline.
Are folks still thinking orbital data centers are an impossibility? Because this is a strong part of the motivation for those.
A data center on a ship out in the ocean is orders of magnitude cheaper and orders of magnitude more capable. If we can't demonstrate a solution to the heat problem, in-orbit DCs that contribute meaningful compute are a fantasy and nothing more.
Solar power in space is potentially very cheap; solar on a ship is not.
The heat problem has demonstrable solutions that people who are physics challenged can't seem to get themselves to accept.
> Solar power in space is potentially very cheap
Only if you ignore the cost of getting in there
You think the heat problem has solutions in space? Good luck! Unfortunately this is a harsh mistress.
https://en.wikipedia.org/wiki/Stefan%E2%80%93Boltzmann_law
https://en.wikipedia.org/wiki/Radiative_cooling
You don't in fact have viable solutions to this problem when orbital launch prices are above like 100USD per ton, probably needs to be closer to 20USD per ton.
What's the solution to the heat problem?
I choose to accept reality. Until someone can do it, it hasn’t been done. Once that happens, this will cease being vaporware. The bubble is filled with vapors and those who drank the koolaide can’t seem to get themselves to accept.
OTOH wave power should potentially be very cheap in the middle of the ocean.
Solar power in space is not cheap. It's efficient. A solar panel in space produces 4-10X as much power as one on land.
Which would be relevant if solar panels or land were scarce and expensive. Neither are. It's easier and cheaper to put 4-10X the number of panels on land.
> The heat problem has demonstrable solutions that people who are physics challenged can't seem to get themselves to accept.
Go on then, what are they? Demonstrable, you said. Show us some demonstrations please.
I've yet to hear of any solution to the heat problem, only vague handwaves that it will be solved. But it's the single largest and most challenging obstacle (amongst a set of large, challenging obstacles)
My searches aren't turning anything up, would you mind sharing what you know about solutions to the heat problem or point me towards a resource if can learn from?
It still doesn't make any sense: if they're not going to use bodies of water to cool them in space, they can not use a body of water to cool them in the middle of nowhere. The same with the solar panels
The water moves the heat from the DC to some near-infinite sink (ocean, river, atmosphere if evaporative, whatever.) In space it's quite (very) hard to move the heat anywhere. So water cooled ocean DC is more realistic than an orbital DC.
If they were talking about putting them on the moon I'd get it; use the moon itself as a giant heat sink. Latency isn't ideal but for applications where that's not a problem (like AI; 3s round trip isn't meaningful in the context of a chat session) then why not.
Not impossible but I'm still curious how they're planning to deal with heat.
I'd also wonder if the folks who don't want to live within 3 miles of a data center are OK with the data center flying over their head
No they aren't because it was never about the externalities of the data center itself (for most - it was for a few people who actually were screwed by the noise, for example). Instead, it's because most of this crowd are anti-AI luddites who hate that change is happening too fast for their minds to comprehend or adapt to.
They stand in the way of a nearly limitless number of positive outcomes from this technology and directly have blood on their hands that they wipe away with thought terminating cliches like "precautionary principle"
Or look at it this way, a massive change has been forced on the public through the decisions of a small number of people. Basically the whole "Accept your fate, it is inevitable" backfired spectacularly. The people were robbed of their agency, their democratic process. This is just the democratic will reasserting its agency.
The modern version of democratic will is a sudden need to ban everything because you saw some outrage articles on social media.
I think that ignores the scale and speed of what is happening with the buildout. That it's gotten to the point that states are putting in moratoriums, thats pretty incredible. I've never seen political government forces mobilize this quickly except in some sort of emergency situation.
That seems like a lot of conjecture based on very little evidence. The article doesn't mention why they're against data centers.
The obvious alternative explanation is that people are very worried about the externalities of a project as massive as putting data centers in space.
Anyone reading between the lines understands datacenter deployment has little to do with datacenters and a lot to do with kneecapping AI development.
If datacenters in space become a thing, you'll hear your neighbor talking about how rockets poison the air.
It depends on the scale. If space data centers become a thing at the scale Musk and others are talking about (millions of space data centers) then rockets will indeed became a major issue for the atmosphere, as will the space data centers themselves.
The total atmospheric pollution from rockets, even at that scale, will be small compared the total from terrestrial sources but unfortunately atmospheric pollution works like real estate: location matters.
Orbital rockets deposit stuff into the upper stratosphere which is about the worst place you can put. Things there have effects massively disproportionate to the same amount of the same things emitted into the lower atmosphere by terrestrial sources. Some have up to 500x the effect when emitted in the upper stratosphere than when emitted by airplanes or on the ground.
It's been OK so far because in the entire history of rockets there have only been about 7500 orbital launches (which are what matter for the upper stratosphere--see below). With a million space data centers which are expected to each have a useful life of 5 years there would have to be about 2000 Starship launches (@100 data centers per launch) per year just for replacing the ones that have reached EOL. That's after the 10000 launches for initial deployment.
Worse, satellites damage the upper stratosphere on reentry. With a million space data centers that's 200k reentries per year. The cumulative number of reentries so far is around 10k satellites and spent rocket bodies, and around 25k if we count smaller things that we can track like debris.
There have been tens of thousands of suborbital launches that have reached the upper stratosphere but most of them have been sounding rockets which are small enough that the cumulative pollution from all of them is less than a year's worth of current orbital launches. On top of that most of the sounding rockets were for weather monitoring and have been replaced by balloons. Current sounding rocket use is negligible. There have also been ballistic missiles but by the time those are in the upper stratosphere they are ballistic and don't pollute much. Hence it is only orbital launches we have to worry about.
The motivation for this idea is to inflate the price of SPCX.
What’s the situation in China?
Are people there also opposed to data centers?
it seems to me like we care about absolutely anything other than the one really important thing right now
This does not bode well for American prosperity. This is going the way of Europe, complacent and non-competitive. The future belongs to China.
Yeah, but how badly did Silicon Valley fuck this up to unite Bernie Sanders and Steve Bannon in a common crusade?
An 84yr old who doesn't understand the technology and 72yr old political hack? Both of whom are populist and lived in the fringe most of their careers.
Data centers would be less contentious in the U.S. if energy was plentiful and clean. As it is, it externalizes costs onto the public and strains profit-optimized electrical grids while contributing to local pollution and global warming.
The question isn't whether that's true, the question is the magnitude of the effect.
By all accounts, it's completely blown out of proportion. If people genuinely, actually care about energy costs, they can vocally support when calls for setting your thermostat at 78F in summertime go out.
> If people genuinely, actually care about energy costs, they can vocally support when calls for setting your thermostat at 78F in summertime go out.
Or, we can prefer energy being spent on air conditioning, and not on data centers.
Then people should stop using AI...
Agreed.
The same demographic is against building new power plants, especially in their vicinity. There could have been abundant power if Nixon‘s plan to build out nuclear had been allowed to proceed.
Now China does it.