Increased demand drives up prices under a fixed supply regime. Power is, in many places of the US, something approaching a fixed supply regime. This is not a physical requirement of the universe. It is a policy choice that we have made. I would argue that it is not the data centers driving up power prices, it is our policies around how difficult it is to provision new power (or build almost anything at all) in the US.
Congress could fix this problem tomorrow. And in fact there is a bipartisan bill (BAAJA) that would go a long ways (although not all the way) towards fixing it. Now yes even if we fix these policies there would be a time lag on new generation being connected, but there is no reason to leave this problem in place.
The main problem identified in this article is cost allocation, not rate of deployment. Data centers paid only 38% of the grid costs tied to their growth. Speeding up permitting would simply accelerate this subsidy to data centers, with households paying even more, sooner.
Your claim that supply is effectively fixed also seems false. Transmission investment driven by new demand grew 17x in five years. The problem is that capex deployed to serve data centers is being paid for by households, not that it cannot be deployed.
Real-world data also cuts against you. New York and New England are among the hardest places in the country to build energy infrastructure, yet the study found no clear price increase there from data centers.
The bill you cited would fix much of this, but it would do so via provisions forcing data centers to pay for the costs they cause, instead of forcing households to subsidize them.
If your point is “it’s policy,” fine. But the policy causing this is cost allocation, not permitting.
Your comment doesn’t engage with the article’s central finding at all: who pays.
Even if congress "fixed" this problem tomorrow it would still take years to build the increased capacity (as you note). Since we do essentially have a fixed supply regime the fast buildout of data centers we've seen over the last few years has indeed been the cause of increased power prices. I'll also add that several renewable power generation projects had been in the works and this administration killed them (even paying a lot of money to kill them).
There is new data center project 30 miles outside my city that plans to build a 1000MW natural gas power plant and a 500MW biomass plant that will take millions of acres of forest to sustain (to quite the developer it will burn 350-500 truckloads of logs per day).
If this is an example of the kind of buildout we want to streamline, I say no thanks.
One of the perverse things about our current permitting regime is that it's often easier for a company to build a private natural gas power plant connected to their data center than it is for the local utility to build a new wind or solar farm. The former can often be done without any kind of discretionary permit at all, whereas the latter goes through a public process in which residents can complain and sue to block the project.
And both are much easier than getting permits for a high voltage transmission line to another region of the country where it might make more sense to build a wind or solar farm.
My town just found out about the forest service’s intent to do some major logging and the truck traffic is way out of spec to the existing roads to the trees.
The scale the data centers want to operate at maybe way out of spec. to the existing infrastructure.
There are rural parts of Massachusetts that have run out of natural gas and electric delivery bandwidth. The simplest solution is non-grid connected power and heat generation. Doing this requires no real change in regulation since the power sources are stand alone. Local building codes are the only regulations.
Thank you for doing the work of keeping in the know on this sort of topic and notifying us all of the bill. I'll check it out and send a letter to my congressmen if it checks out.
If BAAJA means that data center companies can install jet engine gas turbines to power their needs, people are not going to be happy.
I agree the permitting process has hindered building in America, but I'm not convinced the current regimes in charge can craft legislation that actually helps people. America keeps voting for something new because the last attempt hasn't changed the status quo.
NYT has an op-ed today encouraging BAAJA, but all the comments are ripping it for bias. Notably it seems to strip away protections for historic sites and takes away rights from Indian tribes. Imagine running HVDC lines through tribal lands, over the Grand Canyon or other historic landmarks - things that would be prevented today that would be legal tomorrow.
I haven't read it, and I do support some level of expediency with grid buildout, but it sounds like it opens the door to ruining cultural landmarks, no prevention of abuse for fossil fuel purposes. Finally, and I mean this honestly: If the administration supports it, it probably is something worth pausing to consider.
I find their methodology misleading, though. They're basically taking the cost of transmission infrastructure upgrades and then assigning the cost of those upgrades proportional to peak energy demand of the customer base, not the cost of the infrastructure upgrade.
So if, for example, they estimate that residential customer demand will grow by 10%, then they suggest residential customers should only be paying for 10% of transmission infrastructure upgrades.
That's not how infrastructure upgrades work, though. Connecting 100,000 new residential addresses and areas can cost more than a couple data centers that are close to the power station.
The problem is they don't have access to, or at least they didn't look very hard, to find the actual data about transmission line upgrades and how much data centers actually pay to get connected. Contrary to what I see in a lot of the comments here, big customers do actually get billed for their infrastructure upgrade costs after utilities do their analyses. This paper is ignoring the utilities' analyses and substitutes their own highly simplified proportional-to-peak-demand estimates, and that's how they get their headline number.
One very surprising discovery was that they show data center construction incentives to actually be a good deal:
> State incentives for data centers have provided returns on the investments, if inconsistently, the researchers found. Some states, such as California, reaped $180 in worker income for every dollar spent, while Virginia increased income $2 for every dollar invested. The researchers determined that the presence of an existing, qualified workforce was a stronger indicator of high returns than the amount of incentive spending.
That's not what I expected to read.
This is a topic where details basically don't matter for public opinion, though, so I'm not sure it matters. It aligns with what people
I struggle to understand why data centers are seen as a problem rather than an opportunity. Cities and counties can charge a high property tax to data centers and the residents can benefit. They might even give residents a dividend from the data center property tax. What's wrong with that plan?
The article discussed this... data centers don't pay for the infrastructure upgrades they need to operate - the majority of that cost gets foisted on the public. They also drive up the wholesale cost of electricity, which means we all pay more for it.
Both are solvable with politicians who can be held accountable. But, in many locations, they aren't accountable and can only be trusted to do what's in the interest of their largest donors.
That's all ignoring the various other externalities... aesthetics, noise, pollution.
But, 100%, if a county/town has an underutilized industrial zone, they should build data centers there. But, it also appears data center developers prefer undeveloped plots. Again, also solvable, but somebody needs to do so.
This is a regulation problem, not a datacentre problem.
In Aus, when anyone makes an application to connect to the electricity network, you get charged for ALL upstream upgrades to service your load. That might mean new powerlines, new substations upstream etc. That both shifts the load onto the DC operator AND moves them to areas where spare capacity is available.
Similar thing applies for a water connection.
Datacentres still technically drive up spot price for electricity but with the amount of new generation coming online that will even out over time. We're building DCs like crazy here and the average spot price is actually falling.
If a significant part of the build-out is hurting people due to corruption, what are you actually asking for? Should protesters always make sure to add a footnote mentioning that they're only protesting due to the significant likelihood of negative effects due to corruption?
Because if you're asking for them to refocus from data centers to corrupt politicians, I can confidently predict the outcome, and I believe so can you.
USA appears to have weird inversion from how things could be done, so obvious logical benefits of building datacenters are flipped into negatives - through tax breaks, incoherent infrastructure spending (industrial clients should be effectively subsidizing the infrastructure!) on top of electrical grid that might benefit from being taken behind the barn.
Once you get rid of "USA is a specially broken place", you're truly left only with noise pollution, which is workable to deal with and its costs can in fact be taken by developer. Hell, existing industrial zones or ex-factory land is normally preferred elsewhere.
It really is a bizarre place to live... it's both the richest nation in the world, but also uniquely broken in many really sad and confusing ways.
Pretty sure unlimited campaign spending (much of it without attribution via dark money) is a major contributor. While not legally bribery, it sure does look like a duck. Not to mention the actual bribery (747s and weddings for Trump being the most obvious examples).
> USA appears to have weird inversion from how things could be done
The US has a corruption problem, which is not exceptional - I don't know what place you're comparing the US to that doesn't give the rich money and the right of way where and when they want.
People say it's a "race to the bottom" without realizing what that implies. It implies that data centers will simply get deals in localities that will not benefit from data centers if other places demand that they at least pay for themselves. Why would that happen? It happens because it will personally benefit the people who have been charged with making the decision.
> Once you get rid of "USA is a specially broken place"
Again, compared to who? You think they'll have an easier time in Europe? You think if they want to build them in Peru that the residents will benefit?
I think you are getting at something here. The benefit piece. The proximity to community. But the society in the US is not structured in a way to make this happen. Municipality is basically a bad word at this point. But imagine localized network, compute, green energy, battery etc.
I agree that datacenters should be treated like industrial sites. Most people wouldn't want a micro oil refinery or machining factory put up 300 meters from their house. Regulate datacenters like we would a factory. Require them to account for their inputs and their waste outputs, and prohibit them if noise levels are too high or if a concentration of them affects the local climate through heat waste.
Any municipality that tries to do this will not get a DC built, and the one next door will promise a 3 year tax abatement or other special privileges to get it instead. It's a race to the bottom thing.
You need a coordinated policy to encourage everyone to tax them extra, or to require them to put down money for new electrical infrastructure or so on before they break ground. It would be good to have federal action on that.
Why wouldn’t they search for communities that would charge them the least? And in doing so, we have messed up incentives on both sides of the aisle.
Edge compute should be a federal build out IMO. It is considered a matter of national security, right? I could be wrong, so I'm open to discussion.
We build data centers everywhere of all sizes and shapes, they’re part of US Dept of Energy. Companies and individuals can operate them, plug into them like self-storage unit business model.
They would be powered by advanced nuclear, solar, etc. with loads of build out across the country (energy transmission, maintenance, etc.).
Why do we need to have power generation at the source (yes, it is cheaper), when we could do a large scale jobs and infrastructure project out of it that would keep folks employed for GENERATIONS! The demand never ends.
It might open up a lot of opportunities to have that level of localized edge compute, and historically speaking our computing and energy requirements have only ever grown. Give as many people (and communities) a stake into the future!
Private businesses benefit, since they build new things on top of this infrastructure. People benefit, because maintaining the infrastructure is a cost that's rolled in. Everyone benefits from it directly or indirectly (end users).
US Gov could print $1T+ bonds for this, and they would actually pay out in far more ways IMO.
> Why wouldn’t they search for communities that would charge them the least?
They would and they should and that's efficient and good. The problem is corruption. If I go to a sandwich shop and tell them that $12 for a ham sandwich is too high, I can go from sandwich shop to sandwich shop, and eventually I will find one at a quality I can tolerate at a price I'm willing to pay (let's say $6.)
But with these corporations and billionaires, we assume that if we ask them to pay $12 for a ham sandwich, or even $6, they'll keep moving until they find a shop that will pay them $12 to take a sandwich. It's simply corruption. What they do is the equivalent of paying the cashier $3 to give them the sandwich and $12 from the till. They're shopping for the cashier that is willing to do it.
Taxes would be great. If a giant new datacenter meant more money in the coffers for schools, roads, and everything else that is funded by local taxes, then that sounds like something that some places would really like to see.
But that's not how it rolls. At least around me, there instead seems to be a pattern of cities and counties of giving datacenter builds tax abatements of multiple decades in length.
And unlike something like a hospital or a steel mill or even a warehouse: After the local service industry and trades boom from the initial construction is over with, there's not much else coming out of these places. They don't provide very much in the form of long-term employment positions.
And beyond that, it's not like we know what the shape of information technology will be 30 years from now. It's certainly shaped very differently today than it was in 1996.
(And yeah, I know. If there were no tax incentives for datacenters in Location A, then they'd just build them in Location B instead. I don't have a problem with this at all, but some folks seem to find that result to be fundamentally unconscionable.)
Skill issue. This is just some of Loudoun County's report on the positive tax impact of data centers.
> Data center revenue has also enabled the Board of Supervisors to fund the needs of Loudoun County Public Schools, including funding budget increases for the schools amounting to tens of millions of dollars year-over-year. This revenue source has also enabled Loudoun County to meet the public safety needs of the community, funding resources for the Sheriff’s Office and Loudoun County Fire and Rescue. Revenue from data centers has also funded many programs and services provided by Loudoun County to the community, including libraries, parks, recreation and community centers, and critical services in the areas of child protection and mental health. The Board of Supervisors has also funded capital improvements, such as schools, roads, bridges, and a broad range of facilities. All of these expenditures were made possible while simultaneously lowering the real property tax rate for homeowners.
tax breaks on top of subsidized infrastructure as the utility companies now invest more in infrastructure and raise the prices for all preexisting clients to pay for it.
Markets based on supply and demand all work like this. Increased demand results in the market price increasing.
Electricity demand eventually becomes fuel demand, which has the same thing: more demand results in increased prices for fuel. Since fuels are broadly substitutable in electricity generation, all fuels increase in price when there's (significant) additional demand for electricity and that tends to increase prices of everything with a significant energy component.
When demand increases and supply remains static, prices go up.
Supply capacity improvements are incentivized by this price increase.
(If the incentive to improve supply capacity were present at the current price level, then those improvements would already be underway regardless of whether the world had ever seen a datacenter or not.)
Yes it's a huge job creation/economics boon as well. The data centers are injecting a lot of money into local economies (many of which have been shrinking) and providing skilled jobs.
The actual issue is that they don't. A 100 MW steel mill might employ ~1000 people and contribute ~$1 billion/year to the economy. A 100 MW data center doesn't need that many people even during construction, and its economic contributions are attributed to the location of corporate offices. As far as local people are concerned, any other use for the electricity would be more valuable than a remotely owned data center.
Got any examples of 100 MW steel mills and 100 MW Datacenters that you are comparing here? Would love to see some real side by sides if you actually have any data?
>Yes it's a huge job creation/economics boon as well
Is it though? Maybe the construction itself but after that running a data center doesn't need that many on-premise people, like it can be in the two digit range which is not a lot at all. And even then on-premise people won't be the highest paid engineer jobs but rather security and technicians.
Two digit range is a lot of people in a rural area with a couple thousand residents. That's probably a double digit % increase in jobs in that community.
Datacenters certainly provide many construction jobs during that phase of the project.
During normal operations though, you would have a handful of technicians, probably a facilities manager and frequent contractor visits, probably onsite security.
The facilities contractors may come from far away for maintenance projects, so maybe you help support local hospitality businesses.
Most of the highly paid work at the datacenter will happen remotely.
Every so often, there will be a hardware refresh, but you probably bring in people from elsewhere to do those, if it doesn't make sense to have the onsite techs do it little by little.
Certainly, there's a lot of equipment installed, and any locality with property tax on business equipment should get value from that. But otherwise, almost any other light industrial business would have more employees.
It is actually terrible at job creation. Data centers don't create many jobs at all, and you could argue through second and third order effects that they are job destroyers.
Data centers simply don't require very many people to operate.
Job destroyers where? Maybe in the glass towers of NYC and SF and elsewhere. Creating even a few dozen jobs in rural communities is a massive improvement and boost to the local economy.
Sure, I would. In the scope of giant corpo entities, Costco seems like they'd be a decent-enough neighbor to have. They keep short hours: The earliest they're open for regular knuckle-draggers and ruffians is 9:30AM. They're never open after 8:30PM. On weekends, they wrap up by 6:00 or 7:00PM.
Last time I went to Costco, I parked over by the edge of the lot by fences with single-family homes behind them.
It was a busy time of day, but it seemed quiet enough. I remember thinking to myself, "Self, if I lived there, then the first thing I'd do is install a gate in that fence so I could just walk over to Costco."
And I've certainly lived in much noisier neighborhoods with far fewer opportunities to walk around and conduct commerce. I'll take the noise of Costco parking lot traffic over that of a congested 5-way city intersection any day.
Living on a steady diet of $1.50 all-beef hot dogs (with a Coke (with refill!)) probably wouldn't be the healthiest way, but I think I'd find a balance somehow.
Actually, yes, I do. That would be very convenient. Then again, I'm a city dweller. My grocery store is one block over, I literally just walk there with a fold-up wagon.
That's exactly what the newly elected Democrat congressmen are proposing. Right now, it's the opposite where they get tax breaks, reap resources (electricity, water), while providing zero benefits to the local communities (jobs or funding).
the amount of power needed by datacenters is nothing compared to generation gap. These stories cover up the root cause: in nearly every market, no baseload generation has been added for 15+ years, and is generally declining.
Rates aren't going up because of datacenters. They are going up because no investment has been made for more than 15 years, and we have been offsetting the "crunch" with creative workarounds.
False analogy. Data centers are consumers not producers. They compete with humans for infrastructure capacity. Companies need to foot the bill for improving infrastructure not humans
Yes, absolutely. What do local residents get out of the data centers?
I remember earlier this year somebody posted the results of a large survey that concluded most people would rather have a nuclear power plant in their backyard than a data center. Of course everybody would rather have nuclear power plant in their backyard because then they get cheap, clean, reliable electricity. They get nothing from the data center except competition for basic utilities.
I also remember how the person couldn't believe the results and found them astonishing. WTF. This just reinforced for me how so many people in technology are either narcissistic, short sighted (exceptionally naive), or have some sort of qualified mental health defect.
>Yes, absolutely. What do local residents get out of the data centers?
Jobs in places where there usually aren't any. Some of these places are nearly dead or dying towns/communities and they're getting a huge jolt of new economic activity. Once the data centers are built there will be a steady base of new workers/jobs to maintain them.
How many long term jobs for locals do they create after construction? My experience with similar booms in fracking is that you get a lot of out of town workers and some local jobs catering to them for 2-3 years and then it's pretty quiet and the money slows to a trickle once they've built up infrastructure.
DCs need some guards, some highly trained technicians, and that's mostly it, right? maybe 20-30 jobs per? That's not exactly opening up a new major employer like a factory or steel mill.
If you're facing local protests, my suggestion is offer to fund their schools for 5 years with a bond or something like that. Put the company name on a new power plant. It won't take up most of the cost of building and it'll probably get town hall to approve the zoning for you.
That many employees in a community of a few thousand or even 20 thousand is a massive boost to the local economy. Some places these things are being built have had declining populations for decades and haven't had new jobs in as many years as well.
They aren't building steel mills or factories in too many places these days - why do people keep comparing these to data centers as if it's even an option?
All of the data centers I've worked with have had technicians onsite 24/7. There's just a lot of stuff that sometimes needs doing that's easier with someone there.
I'm sure they pull in from outside the area for large jobs like major equipment refreshes or major network changes. But for day to day stuff (and there's always day to day stuff), it's just a couple people. Some days, they're very busy and my lowpri tickets take a long time to get worked, other days my tickets get completed in minutes.
I know some companies are driving towards less intervention with racked hardware and will wait to disturb a rack until several units have failed, but you still need someone to work on those when the hit the threshold. And it's nice to have someone onsite for simple things like replacing a failing cable.
This has been repeatedly debunked. Data centers provide temporary construction jobs, but those jobs are rarely local unless built in a major metropolitan area. Otherwise the data centers are almost completely automated. The Facebook data center near where I live is absolutely massive and employs about 150 people. Its 5 buildings each about a 1000ft long and 2 or 3 stories, if there is nothing underground.
150 employees in a community of a few thousand or even 20 thousand is a massive boost to the local economy. Some places these things are being built have had declining populations for decades and haven't been new jobs in as many years as well.
> Jobs in places where there usually aren't any ... a huge jolt of new economic activity.
This is the incorrect analysis because it is not what is happening. If you can understand this, you can start to understand the resistance. Data centers are a tangible proxy to Ai and the Oligarchy. They look like the globalization, financialization, and deregulation that wrecked the economies they had prior - on steroids.
my understanding is that there are cases where it is not as bad, but by and large - and through history (corporate tax break attractions) - this has not been the case, outcomes are very situational
How many jobs vs how much does the price of power go up? And are they good paying jobs or are they the equivalent of being a 3PL order packer but for swapping drives?
There's a reason lots of data centers are built around airports (already noisy) or in places that can't support other types of development for some reason.
Many local governments have failed in their responsibility to constituents in approving placement of DCs that should or could have better local projects like housing or retail.
Datacenters are getting sweetheart tax deals, and likely will pay way less in taxes than the people do because in the great US of A corporations can simply not pay taxes.
Do you have another argument that doesn't rely on trickle-down economics?
Yes, the whole point of giving them a reduction in future taxes is to encourage development in otherwise undesirable land.
You would never build residential development at the threshold of an airport for obvious reasons. So the remaining options are either leave the land undeveloped and get nothing, or build something that doesn't need quiet to generate tax revenue.
I'm not advocating for or against specific tax agreements because that is between the company and local government, who represent the people.
I looked at the linked page of corporate tax rates. Scroll down to the next graph showing corporate tax revenues. US corporate tax revenues are insignificant compared the other countries in the graph.
What they get out of not having the nuclear power plant is not getting unexplained cancer in 10 or 20 years.
You can tell them as much as you want how safe nuclear power is. You may even be right. But after Fukushima, and Three Mile Island, and Chernobyl, they don't believe you that it will reliably stay safe forever.
The trouble w/ nuclear power plants isn't that the public hates them, it's that the bond market hates them. Even the sticker price is a little bit too expensive but it usually seems to take several times more money and time to build one than anticipated.
Wind is a thing, batteries are a thing, hydro is a thing...
Maybe all that's still not going to cover you in all cases, but...I'm not anti-nuclear by any means, but my understanding is that the upfront costs are very high and take a long time to recoup even in the best of cases. If it's only used as a backup for when cheaper renewable energy can't be generated or bought I'm not sure if the economics will ever work out anymore.
There's a different between "a thing" and a complete system.
They're still expanding hydro in China because they can tell 2 million people to take a hike and there is nothing they can do about it. I remember reading about how the Glen Canyon dam was going to silt up in 40 years or so back in 1990 and... it did!
A real price for solar energy when you consider batteries for overnight/Dunkelflaut and overbuilding for winter could be 2-3x what people are quoting now... Right about in AP1000 territory.
And yeah, what will really happen is that energy markets will be built around the need of gas turbine peakers who will wind up making almost all the profit in electricity generation even if they are only making 1% of it. By 2050 renewables might be seen as a failed technology the way nuclear was circa 1990, but it might not matter much because India will have a heat event that causes a few million mass causalities and they might decide to use their nuclear weapons to put an end to the fossil fuel age.
... People don't pay directly for either nuclear power plants or datacenters. The closest they come to affecting this is by passing and enforcing zoning regulations.
In any case, most of us would rather have one beating than six, but that doesn't mean we want one beating.
I’m lucky enough (/s) to have a regulated, for-profit power company as the only option for electricity. They made $430M in profit this past year and are asking the corporation commission to raise our rates “only” 14% this year. Meanwhile, the cost of electricity they pay to purchase from residential solar is approaching zero.
Unfortunately, american idiocies in building datacenters are exported world wide as "memes" resulting in people thinking it's truth about datacenters, instead of being truth about broken politics and tax policies of USA
Here's an interesting coincidence. During 2010-2023, data centers growth was all in crypto farms, a truly ridiculous massive waste of energy. However, that growth played against the West. It diverted capital into ruinous tulip bubbles. It created an alternative to the Western-led financial system and trivialized bypassing sanctions. Yet it got fairly little mainstream coverage.
Now there's a tech race. Data centers are a necessary evil for the West to win it. And suddenly they're getting massive negative coverage. Just today I saw articles about both the far left and far right prioritizing the issue.
At first blush I'd say the difference in data center growth is significant between the two events, plus there's a much more massive capital inflow to AI datacenters, plus crypto has obvious other things to rail against before you get to noise and water impact.
I'd want to see a lot more evidence for an actual conspiracy than this.
I asked a chatbot to compare the energy expended in mining Bitcoin and on generative AI thus far, and it came up with 900 TWh for Bitcoin and 450 TWh for AI. Granted it might be biased in favor of AI.
The fact that made up issues like water use or waste heat, complete with faked research papers, are being used to push the issue? It does make one suspicious.
There is a lot of the broader anti-AI sentiment, so it's possible that the entire thing is just "people want to hear AI bad, so there's market in saying AI bad" - with media and politicians both recognizing the opportunity. But it wouldn't surprise me at all if some of it turned out to be coordinated influence campaigns.
thats what it kinda feels like to me. My intuition tells me the average non tech person has a very worried view of the future due to AI-- I mean, what industry out there says their industry has a non zero chance of ending life as we know it! So that in mind, people look for reasons to shore up their concerns. Datacenters = AI, therefore they are bad for all these reasons-- Reasons which could be applied to many industries.
I agree with all your points. As with the rest of the anti-AI movement, I think the key is in the likely consequences. Slowing down the buildout and regulating AI in the West won't affect it elsewhere. At best it's tilting at windmills, so I think it's reasonable to suspect the worst.
Congress could fix this problem tomorrow. And in fact there is a bipartisan bill (BAAJA) that would go a long ways (although not all the way) towards fixing it. Now yes even if we fix these policies there would be a time lag on new generation being connected, but there is no reason to leave this problem in place.
Your claim that supply is effectively fixed also seems false. Transmission investment driven by new demand grew 17x in five years. The problem is that capex deployed to serve data centers is being paid for by households, not that it cannot be deployed.
Real-world data also cuts against you. New York and New England are among the hardest places in the country to build energy infrastructure, yet the study found no clear price increase there from data centers.
The bill you cited would fix much of this, but it would do so via provisions forcing data centers to pay for the costs they cause, instead of forcing households to subsidize them.
If your point is “it’s policy,” fine. But the policy causing this is cost allocation, not permitting.
Your comment doesn’t engage with the article’s central finding at all: who pays.
Even if congress "fixed" this problem tomorrow it would still take years to build the increased capacity (as you note). Since we do essentially have a fixed supply regime the fast buildout of data centers we've seen over the last few years has indeed been the cause of increased power prices. I'll also add that several renewable power generation projects had been in the works and this administration killed them (even paying a lot of money to kill them).
If this is an example of the kind of buildout we want to streamline, I say no thanks.
And both are much easier than getting permits for a high voltage transmission line to another region of the country where it might make more sense to build a wind or solar farm.
The scale the data centers want to operate at maybe way out of spec. to the existing infrastructure.
https://www.iberkshires.com/story/84456/Town-Residents-Conce...
https://nicholasinstitute.duke.edu/publications/rethinking-l...
I agree the permitting process has hindered building in America, but I'm not convinced the current regimes in charge can craft legislation that actually helps people. America keeps voting for something new because the last attempt hasn't changed the status quo.
I haven't read it, and I do support some level of expediency with grid buildout, but it sounds like it opens the door to ruining cultural landmarks, no prevention of abuse for fossil fuel purposes. Finally, and I mean this honestly: If the administration supports it, it probably is something worth pausing to consider.
I find their methodology misleading, though. They're basically taking the cost of transmission infrastructure upgrades and then assigning the cost of those upgrades proportional to peak energy demand of the customer base, not the cost of the infrastructure upgrade.
So if, for example, they estimate that residential customer demand will grow by 10%, then they suggest residential customers should only be paying for 10% of transmission infrastructure upgrades.
That's not how infrastructure upgrades work, though. Connecting 100,000 new residential addresses and areas can cost more than a couple data centers that are close to the power station.
The problem is they don't have access to, or at least they didn't look very hard, to find the actual data about transmission line upgrades and how much data centers actually pay to get connected. Contrary to what I see in a lot of the comments here, big customers do actually get billed for their infrastructure upgrade costs after utilities do their analyses. This paper is ignoring the utilities' analyses and substitutes their own highly simplified proportional-to-peak-demand estimates, and that's how they get their headline number.
One very surprising discovery was that they show data center construction incentives to actually be a good deal:
> State incentives for data centers have provided returns on the investments, if inconsistently, the researchers found. Some states, such as California, reaped $180 in worker income for every dollar spent, while Virginia increased income $2 for every dollar invested. The researchers determined that the presence of an existing, qualified workforce was a stronger indicator of high returns than the amount of incentive spending.
That's not what I expected to read.
This is a topic where details basically don't matter for public opinion, though, so I'm not sure it matters. It aligns with what people
Both are solvable with politicians who can be held accountable. But, in many locations, they aren't accountable and can only be trusted to do what's in the interest of their largest donors.
That's all ignoring the various other externalities... aesthetics, noise, pollution.
But, 100%, if a county/town has an underutilized industrial zone, they should build data centers there. But, it also appears data center developers prefer undeveloped plots. Again, also solvable, but somebody needs to do so.
In Aus, when anyone makes an application to connect to the electricity network, you get charged for ALL upstream upgrades to service your load. That might mean new powerlines, new substations upstream etc. That both shifts the load onto the DC operator AND moves them to areas where spare capacity is available.
Similar thing applies for a water connection.
Datacentres still technically drive up spot price for electricity but with the amount of new generation coming online that will even out over time. We're building DCs like crazy here and the average spot price is actually falling.
Because if you're asking for them to refocus from data centers to corrupt politicians, I can confidently predict the outcome, and I believe so can you.
Once you get rid of "USA is a specially broken place", you're truly left only with noise pollution, which is workable to deal with and its costs can in fact be taken by developer. Hell, existing industrial zones or ex-factory land is normally preferred elsewhere.
It really is a bizarre place to live... it's both the richest nation in the world, but also uniquely broken in many really sad and confusing ways.
Pretty sure unlimited campaign spending (much of it without attribution via dark money) is a major contributor. While not legally bribery, it sure does look like a duck. Not to mention the actual bribery (747s and weddings for Trump being the most obvious examples).
The US has a corruption problem, which is not exceptional - I don't know what place you're comparing the US to that doesn't give the rich money and the right of way where and when they want.
People say it's a "race to the bottom" without realizing what that implies. It implies that data centers will simply get deals in localities that will not benefit from data centers if other places demand that they at least pay for themselves. Why would that happen? It happens because it will personally benefit the people who have been charged with making the decision.
> Once you get rid of "USA is a specially broken place"
Again, compared to who? You think they'll have an easier time in Europe? You think if they want to build them in Peru that the residents will benefit?
You need a coordinated policy to encourage everyone to tax them extra, or to require them to put down money for new electrical infrastructure or so on before they break ground. It would be good to have federal action on that.
Edge compute should be a federal build out IMO. It is considered a matter of national security, right? I could be wrong, so I'm open to discussion.
We build data centers everywhere of all sizes and shapes, they’re part of US Dept of Energy. Companies and individuals can operate them, plug into them like self-storage unit business model.
They would be powered by advanced nuclear, solar, etc. with loads of build out across the country (energy transmission, maintenance, etc.).
Why do we need to have power generation at the source (yes, it is cheaper), when we could do a large scale jobs and infrastructure project out of it that would keep folks employed for GENERATIONS! The demand never ends.
It might open up a lot of opportunities to have that level of localized edge compute, and historically speaking our computing and energy requirements have only ever grown. Give as many people (and communities) a stake into the future!
Private businesses benefit, since they build new things on top of this infrastructure. People benefit, because maintaining the infrastructure is a cost that's rolled in. Everyone benefits from it directly or indirectly (end users).
US Gov could print $1T+ bonds for this, and they would actually pay out in far more ways IMO.
They would and they should and that's efficient and good. The problem is corruption. If I go to a sandwich shop and tell them that $12 for a ham sandwich is too high, I can go from sandwich shop to sandwich shop, and eventually I will find one at a quality I can tolerate at a price I'm willing to pay (let's say $6.)
But with these corporations and billionaires, we assume that if we ask them to pay $12 for a ham sandwich, or even $6, they'll keep moving until they find a shop that will pay them $12 to take a sandwich. It's simply corruption. What they do is the equivalent of paying the cashier $3 to give them the sandwich and $12 from the till. They're shopping for the cashier that is willing to do it.
But that's not how it rolls. At least around me, there instead seems to be a pattern of cities and counties of giving datacenter builds tax abatements of multiple decades in length.
And unlike something like a hospital or a steel mill or even a warehouse: After the local service industry and trades boom from the initial construction is over with, there's not much else coming out of these places. They don't provide very much in the form of long-term employment positions.
And beyond that, it's not like we know what the shape of information technology will be 30 years from now. It's certainly shaped very differently today than it was in 1996.
(And yeah, I know. If there were no tax incentives for datacenters in Location A, then they'd just build them in Location B instead. I don't have a problem with this at all, but some folks seem to find that result to be fundamentally unconscionable.)
> Data center revenue has also enabled the Board of Supervisors to fund the needs of Loudoun County Public Schools, including funding budget increases for the schools amounting to tens of millions of dollars year-over-year. This revenue source has also enabled Loudoun County to meet the public safety needs of the community, funding resources for the Sheriff’s Office and Loudoun County Fire and Rescue. Revenue from data centers has also funded many programs and services provided by Loudoun County to the community, including libraries, parks, recreation and community centers, and critical services in the areas of child protection and mental health. The Board of Supervisors has also funded capital improvements, such as schools, roads, bridges, and a broad range of facilities. All of these expenditures were made possible while simultaneously lowering the real property tax rate for homeowners.
https://www.loudoun.gov/6409/Tax-Revenues-County-Budget
Markets based on supply and demand all work like this. Increased demand results in the market price increasing.
Electricity demand eventually becomes fuel demand, which has the same thing: more demand results in increased prices for fuel. Since fuels are broadly substitutable in electricity generation, all fuels increase in price when there's (significant) additional demand for electricity and that tends to increase prices of everything with a significant energy component.
Supply capacity improvements are incentivized by this price increase.
(If the incentive to improve supply capacity were present at the current price level, then those improvements would already be underway regardless of whether the world had ever seen a datacenter or not.)
Is it though? Maybe the construction itself but after that running a data center doesn't need that many on-premise people, like it can be in the two digit range which is not a lot at all. And even then on-premise people won't be the highest paid engineer jobs but rather security and technicians.
During normal operations though, you would have a handful of technicians, probably a facilities manager and frequent contractor visits, probably onsite security.
The facilities contractors may come from far away for maintenance projects, so maybe you help support local hospitality businesses.
Most of the highly paid work at the datacenter will happen remotely.
Every so often, there will be a hardware refresh, but you probably bring in people from elsewhere to do those, if it doesn't make sense to have the onsite techs do it little by little.
Certainly, there's a lot of equipment installed, and any locality with property tax on business equipment should get value from that. But otherwise, almost any other light industrial business would have more employees.
Data centers simply don't require very many people to operate.
Last time I went to Costco, I parked over by the edge of the lot by fences with single-family homes behind them.
It was a busy time of day, but it seemed quiet enough. I remember thinking to myself, "Self, if I lived there, then the first thing I'd do is install a gate in that fence so I could just walk over to Costco."
And I've certainly lived in much noisier neighborhoods with far fewer opportunities to walk around and conduct commerce. I'll take the noise of Costco parking lot traffic over that of a congested 5-way city intersection any day.
Living on a steady diet of $1.50 all-beef hot dogs (with a Coke (with refill!)) probably wouldn't be the healthiest way, but I think I'd find a balance somehow.
Well, that's not the plan. The plan is massive tax cut, infrastructure built with public money and... that's it!
It's about more than money. People live in rural areas for reasons and data centers are rather antithetical to them.
If you want to know details, listen to those people
Rates aren't going up because of datacenters. They are going up because no investment has been made for more than 15 years, and we have been offsetting the "crunch" with creative workarounds.
In that case, the accurate analogy would be: Our kids will look back and wonder why we haven't built many more of them.
I remember earlier this year somebody posted the results of a large survey that concluded most people would rather have a nuclear power plant in their backyard than a data center. Of course everybody would rather have nuclear power plant in their backyard because then they get cheap, clean, reliable electricity. They get nothing from the data center except competition for basic utilities.
I also remember how the person couldn't believe the results and found them astonishing. WTF. This just reinforced for me how so many people in technology are either narcissistic, short sighted (exceptionally naive), or have some sort of qualified mental health defect.
Jobs in places where there usually aren't any. Some of these places are nearly dead or dying towns/communities and they're getting a huge jolt of new economic activity. Once the data centers are built there will be a steady base of new workers/jobs to maintain them.
DCs need some guards, some highly trained technicians, and that's mostly it, right? maybe 20-30 jobs per? That's not exactly opening up a new major employer like a factory or steel mill.
If you're facing local protests, my suggestion is offer to fund their schools for 5 years with a bond or something like that. Put the company name on a new power plant. It won't take up most of the cost of building and it'll probably get town hall to approve the zoning for you.
They aren't building steel mills or factories in too many places these days - why do people keep comparing these to data centers as if it's even an option?
Do they even need the technicians on full time? I feel like those are flown out if something is wrong and otherwise can just monitor remotely?
I'm sure they pull in from outside the area for large jobs like major equipment refreshes or major network changes. But for day to day stuff (and there's always day to day stuff), it's just a couple people. Some days, they're very busy and my lowpri tickets take a long time to get worked, other days my tickets get completed in minutes.
I know some companies are driving towards less intervention with racked hardware and will wait to disturb a rack until several units have failed, but you still need someone to work on those when the hit the threshold. And it's nice to have someone onsite for simple things like replacing a failing cable.
This is the incorrect analysis because it is not what is happening. If you can understand this, you can start to understand the resistance. Data centers are a tangible proxy to Ai and the Oligarchy. They look like the globalization, financialization, and deregulation that wrecked the economies they had prior - on steroids.
my understanding is that there are cases where it is not as bad, but by and large - and through history (corporate tax break attractions) - this has not been the case, outcomes are very situational
There's a reason lots of data centers are built around airports (already noisy) or in places that can't support other types of development for some reason.
Many local governments have failed in their responsibility to constituents in approving placement of DCs that should or could have better local projects like housing or retail.
Do you have another argument that doesn't rely on trickle-down economics?
You would never build residential development at the threshold of an airport for obvious reasons. So the remaining options are either leave the land undeveloped and get nothing, or build something that doesn't need quiet to generate tax revenue.
I'm not advocating for or against specific tax agreements because that is between the company and local government, who represent the people.
Oh, and by the way, not that you'd care given your aggressive tone, but US corporate tax rate is about in line with OECD rates: https://taxpolicycenter.org/briefing-book/how-do-us-corporat...
You can tell them as much as you want how safe nuclear power is. You may even be right. But after Fukushima, and Three Mile Island, and Chernobyl, they don't believe you that it will reliably stay safe forever.
As a result, electricity prices set a new record. At peak my price per kw/h was 22 times higher than it was in the summer. Twenty two times.
Maybe all that's still not going to cover you in all cases, but...I'm not anti-nuclear by any means, but my understanding is that the upfront costs are very high and take a long time to recoup even in the best of cases. If it's only used as a backup for when cheaper renewable energy can't be generated or bought I'm not sure if the economics will ever work out anymore.
They're still expanding hydro in China because they can tell 2 million people to take a hike and there is nothing they can do about it. I remember reading about how the Glen Canyon dam was going to silt up in 40 years or so back in 1990 and... it did!
A real price for solar energy when you consider batteries for overnight/Dunkelflaut and overbuilding for winter could be 2-3x what people are quoting now... Right about in AP1000 territory.
And yeah, what will really happen is that energy markets will be built around the need of gas turbine peakers who will wind up making almost all the profit in electricity generation even if they are only making 1% of it. By 2050 renewables might be seen as a failed technology the way nuclear was circa 1990, but it might not matter much because India will have a heat event that causes a few million mass causalities and they might decide to use their nuclear weapons to put an end to the fossil fuel age.
People (in the US anyway, ymmv elsewhere) really, really hate datacenters and just vaguely distrust nuclear power (without a lot of conviction).
In any case, most of us would rather have one beating than six, but that doesn't mean we want one beating.
>California hasn't built its bullet train because the state has required 130,000 pages of environmental reports, making absurd demands like:
• counting bumblebees one by one • rebuilding woodrat stick huts • hiring specialized dogs to sniff lizard scat
Not all buildings are built or powered the same way in all countries based on the weather of their location.
American lenses about American Datacenters will be about whats happening in America, not other places.
Now there's a tech race. Data centers are a necessary evil for the West to win it. And suddenly they're getting massive negative coverage. Just today I saw articles about both the far left and far right prioritizing the issue.
I'd want to see a lot more evidence for an actual conspiracy than this.
There is a lot of the broader anti-AI sentiment, so it's possible that the entire thing is just "people want to hear AI bad, so there's market in saying AI bad" - with media and politicians both recognizing the opportunity. But it wouldn't surprise me at all if some of it turned out to be coordinated influence campaigns.