This is something, but it is not enough. I live in Virginia. Virginia has been increasingly importing more and more electricity from West Virginia (coal) and Pennsylvania (natural gas) to power new data centers. Between 2019 and 2023, imports grew 61% [1]. Even though Virginia has been building out solar and offshore wind, between 2023 and 2025 low carbon electricity (renewables + nuclear) actually decreased as a % of the energy mix (28% to 26%) [2]. Virginia won’t commit to building the clean energy generation needed to power the data centers and it won’t build fossil fuel powered generation (which is on paper something I agree with). So it imports the electricity that is both dirty and expensive, and we pay for it. It’s the coward’s way out.
Sometimes is hard to understand what really happens for me. So regular people compete on price with companies? That is wild. Here the industry in general is the first one to get disconnected from power if there is not enough. This also creates an incentive for people with money to solve the power generation problem that they themselves created.
The sooner cities and states start negotiating adversarially with companies that want to set up infrastructure in their state, the sooner we'll all be happier with the way things work.
If you want to be a part of my community, you need to obviously be giving at least as much as you take. And none of this can be theoretical in terms of what you're offering. Theoretical == non-existent.
They already do that. It's just that things like large data centers... actually pay a lot of property taxes. A huge amount compared to their land use. While not really consuming a lot of government services.
A building that just needs you to run a transmission line to it once and then sits there unassumingly while paying a disproportionate amount of property tax is the sort of thing municipalities generally offer significant incentives to get to move in. Incentives meaning they pay millions of dollars in taxes instead of tens of millions, but then the municipality gets millions of dollars in tax revenue instead of nothing when another location offers a better deal.
You're already describing a race to the bottom there IMHO.
Cities or even countries underbid each other to get the golden goose, and then realise they've already bargained away any share of the eggs, and it turns out the goose still shits everywhere.
A competitive bidding process isn't a race to the bottom because it doesn't stop at the bottom, it stops when underbidding someone else is no longer worth it. Which happens when the benefit to the one who wins is still a positive number. It's just not a large number, because otherwise it would still be worth it for someone else to make a better offer.
Obviously it's also possible to make a poor choice and provide more in incentives than you get from having the thing, but in that case they're bidding too low and are only suffering the consequences of their own incompetence.
Your scenario there, bidding too low, is a frequent occurrence from what I see. One often exacerbated by exaggerated promises of revenues or employment from the business which then fail to materialise.
Or they come with undisclosed externalities in the form of noise or competition for energy (see for example various bitcoin mines, IIRC there’s one in Texas that is particularly hated by the locals)
The whole idea of cities bidding for these businesses stinks, IMHO.
There is a propensity for the city that gets the business to be the one that offered too much, because if anyone is offering too much then they "win" because no one competent is going to underbid them. But this is the efficient market hypothesis. The place offering too much buys themselves a lesson in not doing that, the next time they don't offer as much (as do other cities who saw what happened there), and this time the city that gets the business gets a net positive result. Then other cities see that and want it and if any of them are incompetent then they offer too much and suffer the consequences again. The average converges on a small positive return for the city that wins. Or a negative return if too many cities are incompetently governed, but your problem in that case is the incompetent governance rather than the premise that offering something that does bring a net positive return to the city would be a bad choice.
And what is the alternative supposed to be? Cities A, B and C all have different tax rates, other policies, labor force, geographical proximity, etc. Companies will build wherever it's most advantageous. Some town in rural Illinois is at an intrinsic disadvantage compared to coastal cities unless they can offer some countervailing advantage like lower tax rates. The coastal cities may not like that because they want to keep the business even if they're diverting the tax money to cronies or doing nothing to address high cost of living or otherwise governing like schmucks, which is exactly why the competition is good. It's what causes the people in those cities to ask why tax dollars are going to waste and corruption instead of ignoring that because the money is being extracted from a company which is somehow prevented from moving to a less corrupt city.
It's one of the only things that provides governments an incentive to improve efficiency.
I'm not sure what the status is, getting hard to find out amid so much noise & public coverage that's just blathering heads mostly, but, for the longest time & in many places, data centers were getting charged at industrial rates for power, which is intentionally kept super low to encourage industry. The thought that they are paying vastly less as everyone else's prices goes up is a bitter detestable situation. I am nowhere near as anti-data center as most, I want these things actually, but it feels like it's such a strong example of elite revolt, of people being totally paved over by the rich and powerful, who have so effectively managed to buy out the system of governance in so many ways.
There's nuance of course, but in VA, in one county, 37 data centers account for 3GW of power.
They're getting that at those vastly discounted industrial rates. Meanwhile, their utilities are getting ready to foist nearly $20B of works to bring in about that same 2-3GW in from elsewhere because of capacity issues onto all their customers, residential and commercial alike.
"Nuance" is perhaps soft-balling it a bit, ha. Yes, MAE-EAST formerly MAE, the home of the internet, is indeed a bit of a data-center hotbed. https://en.wikipedia.org/wiki/MAE-East
So now every data center will have run their own wires? The city will be covered in transmission towers, wires, substations. To place new transmission wires, they’ll have to cut down some forest or demolish something.
Oh no! Whatever will we do without AI! It'll be like the dark ages of six months ago all over again. Or maybe it'll be like the reaaaallly dark ages twelve whole months ago, back when humans hadn't yet been overtaken by AGI
The power upgrades are required, yes. In some cases people have had their property taken by eminent domain just for power line upgrades. And people wonder why there is so much opposition.
The change here is that the data center will pay the full cost of the upgrade.
This is something, but it is not enough. I live in Virginia. Virginia has been increasingly importing more and more electricity from West Virginia (coal) and Pennsylvania (natural gas) to power new data centers. Between 2019 and 2023, imports grew 61% [1]. Even though Virginia has been building out solar and offshore wind, between 2023 and 2025 low carbon electricity (renewables + nuclear) actually decreased as a % of the energy mix (28% to 26%) [2]. Virginia won’t commit to building the clean energy generation needed to power the data centers and it won’t build fossil fuel powered generation (which is on paper something I agree with). So it imports the electricity that is both dirty and expensive, and we pay for it. It’s the coward’s way out.
[1] https://cardinalnews.org/2025/01/21/virginia-now-imports-mor...
[2] https://lowcarbonpower.org/region/Virginia
Sometimes is hard to understand what really happens for me. So regular people compete on price with companies? That is wild. Here the industry in general is the first one to get disconnected from power if there is not enough. This also creates an incentive for people with money to solve the power generation problem that they themselves created.
Agriculture pays, on average, probably something like 1/1000th of what residential users pay for water.
Does much agriculture use the sanitized municipal water supply?
Its only fair. We shoudlnt socialize costs for things where the profit will be private
Which makes me think of publicly owned data centers. Interesting thought
The sooner cities and states start negotiating adversarially with companies that want to set up infrastructure in their state, the sooner we'll all be happier with the way things work.
If you want to be a part of my community, you need to obviously be giving at least as much as you take. And none of this can be theoretical in terms of what you're offering. Theoretical == non-existent.
They already do that. It's just that things like large data centers... actually pay a lot of property taxes. A huge amount compared to their land use. While not really consuming a lot of government services.
A building that just needs you to run a transmission line to it once and then sits there unassumingly while paying a disproportionate amount of property tax is the sort of thing municipalities generally offer significant incentives to get to move in. Incentives meaning they pay millions of dollars in taxes instead of tens of millions, but then the municipality gets millions of dollars in tax revenue instead of nothing when another location offers a better deal.
You're already describing a race to the bottom there IMHO.
Cities or even countries underbid each other to get the golden goose, and then realise they've already bargained away any share of the eggs, and it turns out the goose still shits everywhere.
A competitive bidding process isn't a race to the bottom because it doesn't stop at the bottom, it stops when underbidding someone else is no longer worth it. Which happens when the benefit to the one who wins is still a positive number. It's just not a large number, because otherwise it would still be worth it for someone else to make a better offer.
Obviously it's also possible to make a poor choice and provide more in incentives than you get from having the thing, but in that case they're bidding too low and are only suffering the consequences of their own incompetence.
Your scenario there, bidding too low, is a frequent occurrence from what I see. One often exacerbated by exaggerated promises of revenues or employment from the business which then fail to materialise.
Or they come with undisclosed externalities in the form of noise or competition for energy (see for example various bitcoin mines, IIRC there’s one in Texas that is particularly hated by the locals)
The whole idea of cities bidding for these businesses stinks, IMHO.
There is a propensity for the city that gets the business to be the one that offered too much, because if anyone is offering too much then they "win" because no one competent is going to underbid them. But this is the efficient market hypothesis. The place offering too much buys themselves a lesson in not doing that, the next time they don't offer as much (as do other cities who saw what happened there), and this time the city that gets the business gets a net positive result. Then other cities see that and want it and if any of them are incompetent then they offer too much and suffer the consequences again. The average converges on a small positive return for the city that wins. Or a negative return if too many cities are incompetently governed, but your problem in that case is the incompetent governance rather than the premise that offering something that does bring a net positive return to the city would be a bad choice.
And what is the alternative supposed to be? Cities A, B and C all have different tax rates, other policies, labor force, geographical proximity, etc. Companies will build wherever it's most advantageous. Some town in rural Illinois is at an intrinsic disadvantage compared to coastal cities unless they can offer some countervailing advantage like lower tax rates. The coastal cities may not like that because they want to keep the business even if they're diverting the tax money to cronies or doing nothing to address high cost of living or otherwise governing like schmucks, which is exactly why the competition is good. It's what causes the people in those cities to ask why tax dollars are going to waste and corruption instead of ignoring that because the money is being extracted from a company which is somehow prevented from moving to a less corrupt city.
It's one of the only things that provides governments an incentive to improve efficiency.
Now this is the pragmatic solution. Very clever
Virginia also has a tax on electricity, targetting data centers. https://www.forbes.com/sites/robertszczerba/2026/07/13/power...
I'm not sure what the status is, getting hard to find out amid so much noise & public coverage that's just blathering heads mostly, but, for the longest time & in many places, data centers were getting charged at industrial rates for power, which is intentionally kept super low to encourage industry. The thought that they are paying vastly less as everyone else's prices goes up is a bitter detestable situation. I am nowhere near as anti-data center as most, I want these things actually, but it feels like it's such a strong example of elite revolt, of people being totally paved over by the rich and powerful, who have so effectively managed to buy out the system of governance in so many ways.
There's nuance of course, but in VA, in one county, 37 data centers account for 3GW of power.
They're getting that at those vastly discounted industrial rates. Meanwhile, their utilities are getting ready to foist nearly $20B of works to bring in about that same 2-3GW in from elsewhere because of capacity issues onto all their customers, residential and commercial alike.
"Nuance" is perhaps soft-balling it a bit, ha. Yes, MAE-EAST formerly MAE, the home of the internet, is indeed a bit of a data-center hotbed. https://en.wikipedia.org/wiki/MAE-East
There's some good books on this rise too. Internet Alley, for example, https://mitpress.mit.edu/9780262516686/internet-alley/
Yes! Yes! Make a lot of data centers but make them build out new power renewable generation to power them.
So now every data center will have run their own wires? The city will be covered in transmission towers, wires, substations. To place new transmission wires, they’ll have to cut down some forest or demolish something.
Oh no! Whatever will we do without AI! It'll be like the dark ages of six months ago all over again. Or maybe it'll be like the reaaaallly dark ages twelve whole months ago, back when humans hadn't yet been overtaken by AGI
Probably pay the power company to put in the required infrastructure upgrades I guess.
The power upgrades are required, yes. In some cases people have had their property taken by eminent domain just for power line upgrades. And people wonder why there is so much opposition.
The change here is that the data center will pay the full cost of the upgrade.