For those questioning this conclusion, 4 of top 10 states by GW of DC under development flipped parties and elected governors that made curtailing DC development or increasing regs part of their platform. That's 130GW of the ~350GW pipeline in the US.
This obviously matters. These states have the leverage. Look at the issues Oracle is having in NM.
This will encourage developers to put these facilities in more human-friendly places. I knew the clock was ticking when xAI started expanding aggressively in suburban Memphis. These facilities are loud, they emit pollution (when power is generated onsite), and they need very little staff. After the initial construction there is zero upside for the local community. They should be located away from population centers.
more likely they'll get built in different countries with more lax regulations. This is probably going to end up like manufacturing where it's all done overseas.
Property taxes are a significant contributor to the purse of many cities and counties, and datacenters are expensive real estate in terms of the assessed valuations. Sales tax also.
For some cities and counties this will be a loss if those developments move.
I would take a wild guess that data centers have a pretty mediocre assessed value, and of course they have even less of an employment base than a typical warehouse or factory.
Not appreciably in most municipality so far, I guess what your saying is nice in theory but isn't being born out in reality in the simple majority of places.
It is not as if data center development is suddenly ended. Democrats have paired up with tech in the past. Tech might want to make friends with those that hold the purse stings. Regulatory capture is still in the cards. Projects on the way are still on the way.
It might be that tech would like some regulations / laws that slow data center development (maybe a good thing) if they think it has just turned into an arms race that has gotten out of hand. Investors might think this also (S&P 500 up ~1.5% as I write this).
Negative sentiment on data centers is bipartisan. It’s incredibly strong right now.
Regulatory capture won’t save the tech industry if they can’t improve public sentiment.
They’ve been talking down to the average person, doom-bragging about the existential crisis of AI to juice investors, and people like Sam Altman have been generally been acting like Bond villains.
I think their shortsightedness has backed them into a corner of a potentially highly regulated environment where building a data center is harder than building a chemical plant.
Yeah, and both their politicans and constituencies were directly betrayed by them. They made their bed and now they must sleep in it. Many of the dems who won made data centers and AI a part of their platfrom.
And while before you could say well money overrules everything, they outspendint Democrats like crazy and still got wiped out. So if you want to keep your job as a politician suddenly more fundraise no longer means more advantage
I am a little hopeful here in Texas. We have needed change for a while. Still, we are talking about politicians and while there maybe be a mandate from voters it is a long and narrow road to solutions that are good for everybody (especially the constituents).
The nice thing about politicians is they can be held accountable by constituents, per the Constitution.
Data center developers, who should be equally distrusted (but seem to get a pass), are not beholden to you.
It will take developers and politicians working together to find a solution, rather than let developers do whatever they can to maximize profits without regulation or public process.
The scale of the market fuckery is incredible. We are trying to install an MR scanner. Chillers are becoming hard to get and on long back order. The supplier says that the units we want are not really the right units for data centres but due to shortages of the units data centres really want, they are picking up alternatives.
It just shows the utter hollowing out of the US industrial base to me.
Everything has been cut to the bone so any blip in demand cannot - at any price - be effectively responded to. Even if you were exceptionally good at being able to double internal capacity, the supply chain all the way from raw materials to skilled labor pool your vendors may need just isn’t there.
This latest datacenter bubble is just the last and most acute symptom of this problem I’ve noted in my lifetime. Hyper-financialization does that to an economy.
The best anyone can do is outbid peter to rob paul.
Consider that semiconductors are the worst case situation for a supply chain planner:
- most capital-intensive manufacturing process ever invented
- lead time for expansion is several years
- demand that is highly uncertain on a timescale of months
- technology that goes rapidly obsolete
Unless one of these factors changes, it essentially guarantees that semiconductors will have the highest price volatility of any manufactured goods. As they have since the 1980s.
This sub-thread wasn't about semiconductors though. It was for chilling plant. Semiconductors I do agree are a somewhat special case - but also one the US ceded it's capability in and largely offshored.
But it's way more than just chillers. It's everything from structural steel, to electrical transformers, copper cabling, optical transceivers, electric generation, skilled workforce in the trades, etc.
Overall this boom has been relatively tiny as far as US industrial booms historically have been, and we cannot ramp up production even if there was the political will to do so.
It's not just datacenters. Look at the military industrial supply chain these days. We can't even scale up WWII era artillery shell production in any reasonable timeframe. Much less modern weapons systems.
I'm sure there was a "spray and pray" attitude, but also some stunning incompetence... Republican PACs spent a lot of money advertising for a candidate who was running in Portland, Maine... in Portland, Oregon, on the other side of the country.
Darn right! Honestly, I really don't understand how anyone can see any value in this, except those at the top. Yes, LLMs make us more efficient in SOME cases, but does that improve life for the average man or even average programmer? Or even the end user, who has to deal with more AI-generated ads and scams now?
The only end benefit seems to be that more people can make more software now, and it's not even clear that this will be a good thing in the long run, considering how much technical debt we're accumulating.
That hundreds of hackers can complete a few neat hobby projects doesn't really count as a serious benefit, by the way.
If you need to start adding exceptions before you've finished your comment then it seems it must not be hard to come up with how someone can find value in it?
For a random non-software example: My mother has an HR job and absolutely loves it for working with spreadsheets, sorting through email, making mail merges, reviewing the policy documents for errors as she works on them, and so on. She spends less time in document hell and more time getting things done or talking to people.
Personally, I find the ability to quickly remove technical debt to be more practically valuable than the ability to quickly create software. A lot of the smaller shops I work with never had the time or money to properly manage, update, or retire their infrastructure. Now many are in a better shape than my own stuff has been because you don't need to be an uber nerd spending a weekend to upgrade an ancient & neglected MediaWiki box (or whatever) anymore.
In the past few weeks, I have used AI assistance to...
...track down a piece of music I wanted to hear again, based on a description of its general vibe and the single line of the lyrics I could remember
...compare the weather resistance between various models of temporary carport
...figure out how to get the last stubborn bolt out of a snowmobile engine I was trying to remove (hidden underneath, behind a covered access panel, it turned out)
...design a costume I want to make, from a vague concept through inspiration pieces to specific details to a complete design sketch I can use to draft the sewing pattern
...apply a voice-over to a piece of video using kdenlive
...find a replacement ATV master cylinder which is cheaper than the OEM part
...fix an annoying string buzz on a mandolin
...locate the procedure for getting the city to plant a street tree in front of the house
...recommend a show I haven't seen which would fit the mood I was in
...make a recipe for a dish I wanted to cook, combining elements from two different variations I found online, modified to accommodate the ingredients I actually had
...evaluate all the second-hand welders available on the local craigslist based on their suitability for a certain project I want to tackle
...plan a last-minute night out in a city I don't know very well
...find replacement glass to fit a ceiling light fixture
...suggest some music not already in my library which would fit in with a certain playlist
...find out how parking works at an airport I was about to visit
...figure out why my motorcycle wasn't starting reliably
...find a plumber capable of solving an issue with an antique clawfoot tub
...learn how to assemble a DIY meshcore radio kit which came with no instructions, by taking pictures of the parts
I could continue, but I believe the point is made.
Overall I do believe it is a productivity win for software. However, the insane amount of money spend on AI (training models, building datacenters and even usage in some cases) is ridiculous. Hopefully that will get lower soon.
From my time at hyperscalers ... realistic for low power traditional DC (lot of regular rackmount CPU server) ... unrealistic for the power and cooling requirements of inference and crypto mining
Many of the votes are not full ban on DC but ban on massive energy dense DC
We can only hope this trend will continue from defense to offense, actively punishing AI companies and their leaders for the harm they have inflicted and continue to inflict on society.
I’m a progressive through and through, but this is the one aspect of the whole platform I don’t like. I am in the build data centers everywhere, make tokens dirt cheap, and work as hard as possible to make models more and more intelligent camp. The only thing I would like to see is that we keep building massive data centers, but mandate that they have to be powered by their own newly built solar, wind, and battery banks. This would let us keep our lead in frontier models, as well as building out brand new sources of green energy. I believe Massachusetts (?) has already passed legislation that accomplishes this.
Unfortunately the hyper scalers decided to aggressively shoot themselves in the foot by following the Amazon HQ2 approach of making cities compete and signing deals in the dark under NDA. They became symbols of grift and corruption and undemocratic strip mining of communities for resources because they figured they could culture war their way out of any consequences. It's very stupid because there's no way this is worth it financially for them given how much backlash it generated. Given how hyped unions were/are about data center construction just showing up as a good neighbor and not being an asshole would have kept public support but somehow they didn't realize how much the generally positive impressions of tech the public used to have were worth in soft power and just how much value they were throwing into the trash by nuking it.
I say this out of genuine curiosity: does Florida even make sense for data centers with the increasing weather patterns there? Feels like every time I read news about Florida, if it’s not political, it’s some natural disaster.
I assume there must be companies doing things like this there but would the big co’s actually want anything there?
I just used Florida for one counter example but Tennessee and Alabama just rolled out the red carpet, as results have come in, too. The usual course of things like this is the corporations will play different jurisdictions off each other to get more and more off taxpayers dime like pro sports stadiums and Amazon heralded HQ2 - thankfully voters in a lot of states mentioned rejected unlimited data centers - but it is not universal. The fight continues.
This obviously matters. These states have the leverage. Look at the issues Oracle is having in NM.
https://dcmap.us/insights/pipeline/ https://www.cnbc.com/2026/09/24/oracle-data-center-force-maj...
For some cities and counties this will be a loss if those developments move.
It might be that tech would like some regulations / laws that slow data center development (maybe a good thing) if they think it has just turned into an arms race that has gotten out of hand. Investors might think this also (S&P 500 up ~1.5% as I write this).
Regulatory capture won’t save the tech industry if they can’t improve public sentiment.
They’ve been talking down to the average person, doom-bragging about the existential crisis of AI to juice investors, and people like Sam Altman have been generally been acting like Bond villains.
I think their shortsightedness has backed them into a corner of a potentially highly regulated environment where building a data center is harder than building a chemical plant.
Yeah, and both their politicans and constituencies were directly betrayed by them. They made their bed and now they must sleep in it. Many of the dems who won made data centers and AI a part of their platfrom.
They might find that tricky after they sold out to right wing causes last election.
With the mandate from voters, hopefully we see less secret development and more open communication about the deals being cut.
Data center developers, who should be equally distrusted (but seem to get a pass), are not beholden to you.
It will take developers and politicians working together to find a solution, rather than let developers do whatever they can to maximize profits without regulation or public process.
Everything has been cut to the bone so any blip in demand cannot - at any price - be effectively responded to. Even if you were exceptionally good at being able to double internal capacity, the supply chain all the way from raw materials to skilled labor pool your vendors may need just isn’t there.
This latest datacenter bubble is just the last and most acute symptom of this problem I’ve noted in my lifetime. Hyper-financialization does that to an economy.
The best anyone can do is outbid peter to rob paul.
- most capital-intensive manufacturing process ever invented
- lead time for expansion is several years
- demand that is highly uncertain on a timescale of months
- technology that goes rapidly obsolete
Unless one of these factors changes, it essentially guarantees that semiconductors will have the highest price volatility of any manufactured goods. As they have since the 1980s.
But it's way more than just chillers. It's everything from structural steel, to electrical transformers, copper cabling, optical transceivers, electric generation, skilled workforce in the trades, etc.
Overall this boom has been relatively tiny as far as US industrial booms historically have been, and we cannot ramp up production even if there was the political will to do so.
It's not just datacenters. Look at the military industrial supply chain these days. We can't even scale up WWII era artillery shell production in any reasonable timeframe. Much less modern weapons systems.
Wonder how much money the burned.
The only end benefit seems to be that more people can make more software now, and it's not even clear that this will be a good thing in the long run, considering how much technical debt we're accumulating.
That hundreds of hackers can complete a few neat hobby projects doesn't really count as a serious benefit, by the way.
For a random non-software example: My mother has an HR job and absolutely loves it for working with spreadsheets, sorting through email, making mail merges, reviewing the policy documents for errors as she works on them, and so on. She spends less time in document hell and more time getting things done or talking to people.
Personally, I find the ability to quickly remove technical debt to be more practically valuable than the ability to quickly create software. A lot of the smaller shops I work with never had the time or money to properly manage, update, or retire their infrastructure. Now many are in a better shape than my own stuff has been because you don't need to be an uber nerd spending a weekend to upgrade an ancient & neglected MediaWiki box (or whatever) anymore.
...track down a piece of music I wanted to hear again, based on a description of its general vibe and the single line of the lyrics I could remember
...compare the weather resistance between various models of temporary carport
...figure out how to get the last stubborn bolt out of a snowmobile engine I was trying to remove (hidden underneath, behind a covered access panel, it turned out)
...design a costume I want to make, from a vague concept through inspiration pieces to specific details to a complete design sketch I can use to draft the sewing pattern
...apply a voice-over to a piece of video using kdenlive
...find a replacement ATV master cylinder which is cheaper than the OEM part
...fix an annoying string buzz on a mandolin
...locate the procedure for getting the city to plant a street tree in front of the house
...recommend a show I haven't seen which would fit the mood I was in
...make a recipe for a dish I wanted to cook, combining elements from two different variations I found online, modified to accommodate the ingredients I actually had
...evaluate all the second-hand welders available on the local craigslist based on their suitability for a certain project I want to tackle
...plan a last-minute night out in a city I don't know very well
...find replacement glass to fit a ceiling light fixture
...suggest some music not already in my library which would fit in with a certain playlist
...find out how parking works at an airport I was about to visit
...figure out why my motorcycle wasn't starting reliably
...find a plumber capable of solving an issue with an antique clawfoot tub
...learn how to assemble a DIY meshcore radio kit which came with no instructions, by taking pictures of the parts
I could continue, but I believe the point is made.
Something that can be trucked in with no negative effects on local populations
Many of the votes are not full ban on DC but ban on massive energy dense DC
> with no negative effects on local populations
Fan noise unfortunately
It's a very successful Chinese influence campaign.
It's not that hard to build these in a way that isn't mendacious, the total moral collapse of the hyperscalers is not something to celebrate.
I assume there must be companies doing things like this there but would the big co’s actually want anything there?