New AI data centers are obviously compute-oriented instead. That means they consume a lot more power, like hundreds of megawatts. For perspective, on average a U.S. household uses only 1.2KW. Because of this, cooling needs are significantly increased, and while water usage is often overblown, noise volume is a major problem. At 100m, it often exceeds the 90dB threshold for hearing loss.
Then there's the fact that millions of dollars of GPUs are inside these things, so security means lots of bright lights at all times of day.
These are just problems with the physical presence of a data center, not to mention the electric bill hikes, land value loss, environmental effects, etc. And there's no obvious positives to one going up. No new jobs. No tax money because it's often elided with incentives from bribing local officials.
With all the negatives and seemingly no benefits, ask yourself who wants these data centers, and why? If the only way to improve AI is using more and more compute like this, what is the end-game?
> At 100m, it often exceeds the 90dB threshold for hearing loss.
> I've read reports from people living near data centers having almost constant 60-70dB noise levels at their houses, which, yeah, it still shouldn't fly.
I absolutely believe that there's leftover noise around the facilities, too, specially low-frequency humming for electrical equipment, but I have yet to see _actual_ reporting demonstrating newer AI-centered datacenters produce even near the amount of noise OP was claiming.
> Noise level ~10m outside a facility is 90dB.
edit: just understood what "Extra zero" meant. The original comment before being edited said 100m and that's what I was quoting.
I still heard almost nothing ~20m from it, but neither I (nor the company who built it) knew what kinds of workloads ran there and I suspect it was mostly traditional workloads.
For someone who knows enough to point out that dB is logarithmic, you seemed to have neglected the fact that sound is inverse-square for distance? Unless these neighbors are _literally_ located 100m outside the data center?
dB should generally be expressed at what distance it was measured, and you'll see this done all over the place.
In this case, you'd reach 70dB at about 1km from the 90dB at 100m source, assuming a spherical cow, but they could also be completely different datacenters.
OP was talking about datacenters in residential areas and quoted the 90db @ 100m value, which indeed is 70dB @ 1km distance, which is absolutely absurd and would make it impossible to live nearby. That's the original comment I've answered.
Then, OP edited the post and said it was 90dB @ 10m, which means 50dB @ 1000m, an okayish value (depending on the type and duration of the noise), but that's not a distance we were talking about before.
I know datacenters in my city that are less than 500m from residential zoning (I should note I'm not in the US and this is not a new datacenter), so I didn't think that the original (distance) measurement sounded absurd, that's why I pointed out how much the (noise) measurement seemed to be out-of-scale.
Anyways, I think OP's measures where just completely made up or misremembered, because I'm any way I've framed the data (and the edited data), it doesn't match my personal observations around the few datacenters I've been to nor the data that's been released.
I did agree about the possibility of low-frequency noises being a huge nuisance, though, as those types of noises feel much worse than random noises at random intervals at higher dB.
It's also not a uniform medium. Sound gets reflected by houses, the ground, resonance, so it can constructively interfere and have a extra few times the expected amplitude. This can make the sound quaver and become really annoying, too.
Are you serious? Is there any data for this? This would make the whole area unlivable
- explicitly only wanted to eval the infrasound noise burden, which is not what the guy above was referring to here (the human audible noise burden; not that this would be pointless to check)
- played back a recording where the car door slamming was way louder than ambient
- showed that the place is pretty far away from residential areas
- did not actually confirm the source of the infrasound he did record
- did not disclose the actual dB SPL figures measured
- kept derailing to talk about other things he finds objectionable, to the extent that it felt like more time was spent on those than on the actual subject of the video.
If the remaining 20 minutes are like this too, then I'm not sure he's a reliable narrator or that this is actually valuable research, at least in this context.
https://blog.andymasley.com/p/contra-benn-jordan-data-center...
Yes, data centers are loud. Fortunately, some time ago humans achieved magical technological breakthroughs called "earplugs" and "ear protectors".
I've spent enough time on city council noise ordnance boards and understand enough about acoustics to say that the professional sound system I own would certainly be audible and annoying to my neighbors half a kilometer away even if it were not appreciably above the noise floor of the surrounding environment.
Folks are great a signal detection and it's very annoying to them to be able to detect an unwanted bass drum or similar, which they can do at pretty long distances. Even distributed noise is still often noticeable to people.
For instance, I live in the literal middle of nowhere and I still encounter real noise pollution from gas well heads and gas processing plants that are several kilometers from my home.
Sure, some temporary jobs during construction, but I bet it’s nothing like a retail store or factory in operation…
As for taxable revenue, how does that help the community beyond property tax assessments?
The tax breaks are well-established in the media, see for example [0].
That’s the root of the problem.
Even if the data centre raises billions in revenue and pays billions in taxes, that isn’t paid to any government entity close to where the building is.
The locals get replaced by machines, the revenue is pocketed far away.
It’s the equivalent of big mining companies “investing” into mine automation in some developing country. The locals do not benefit!
Who then refuse to sell their machines to anyone except some Taiwanese company.
So, a 100MW DC will have about 20-30 people and a 1GW campus will be 200-300 people.
Those numbers can be boosted if the site has a significant non-operational support staff such as security and cleaning. But a significant part of security us heavily automated, to the point where when Meta had a major network failure they couldn't get on-call support staff into the facilities to fix the issue because the security doors were linked to the main network.
This is always kind of funny to me. How's an average US house supposed to provide context for industrial compute?
How much energy does a car factory use? What about a chip fab? A hospital? A battery manufacturer?
What about an apartment of the same physical dimensions as the proposed "data" center?
Any of those measurements would provide meaningful context that would allow the audience to decide what they'd like built in that space if energy consumption is a concern of theirs.
Comparing a multi-million dollar server farm almost deliberately designed to turn electrons into dollars to a US house usually designed to require as little electricity as possible provides...well I can't decide if it's zero or next to zero context. Like, sure, that power powers say thousands of homes, but my own home consumes more energy than is really required to keep me alive. If the concern is efficiency, I'd be just as inclined to abolish new freestanding homes in favor of efficient apartment complexes as I would be to disfavor a data center over a neighborhood. But those up in arms about data center power consumption are not also rallying for fewer lawns, so the argument can't be purely about power consumption.
>With all the negatives and seemingly no benefits, ask yourself who wants these data centers, and why?
What do you mean no benefits? Do you think local governments are signing these contracts and listening to their constituents yelling at them for the fun of it? The company building it profits from compute contracts. The device mfgs profit from purchasing agreements. The city with the land profits via both legitimate and alleged payments from the corp building the center. To suggest that there are seemingly no benefits just because you aren't receiving the benefits is naive. No one's doing all this for free.
>If the only way to improve AI is using more and more compute like this, what is the end-game?
You could have and probably could still say the same about the internet in general. We've been exponentially building out storage, delivery, and compute for decades. If the only way to improve the web is more and more servers, what is the end game?
You could say the same about just about any growth industry in history. If the only way to reduce the price of bananas is more banana farms, what is the end game?
And the answer is the same in every case. There is not infinite demand for AI and the accompanying infra just as their is not an infinite demand for bananas. If their is not an infinite demand, there is no call for infinite supply increase.
Demand for AI will peak, plateau, and crash, over time just as the tulips, beanie babies, pokemon cards, bitcoin, and everything else does. If you're not in to tulips or pokemon, the craze seems stupid. Because the craze seems stupid, you do not feed it. So long as there are people not feeding it (in order to buy bananas or beanie babies instead of AI subscriptions), its growth and presence will be modulated just as any other thing available for purchase.
tl;dr: rough napkin math and basic googling, 20 MW for a car factory, 100 MW for a chip fab, 2 MW for a hospital, and 0.2 MW for an apartment the same size as the data center. Water usage is equivalent to 14 car factories, 52 hospitals, or 18,000 people at home.
Good questions! You made me curious. In the process of looking, I came across this[0] reddit post with a chart comparing various industrial uses. For car assembly specifically, Honda has some nice global data here[1], and their total energy consumption is about 990 MW, split roughly equally between direct and indirect consumption. That's split across something like 60 factories! If I just look at direct energy consumption, that's less than 20 MW per factory.
This[2] has a chip fab at 100 MW. This[3] has an average hospital at 1.7 MW, with the largest use just being heating. I didn't try to look up battery manufacturers. This[4] has a typical apartment at 0.77 W/square foot, which would mean just 0.2 MW for an apartment the same size as the data center in the article.
I think a more interesting distinction is water usage. The linked article says that the Fireball LLC datacenter used 548 megagallons of water in 2025. Assuming that aligns with the "water consumption" rather than "water withdrawal" numbers from Honda, that's over 20% as much as the entire global operations of Honda equivalent to 14 factories (adjust by about a factor of two if we should be comparing to water consumption). Alternatively, it's equivalent to 52 hospitals[5] or 18,000 people in their homes[6].
[0]https://www.reddit.com/r/aiwars/comments/1tes0rr/ai_data_cen...
[1]https://global.honda/en/environment/initiatives/data/
[2]https://www.mckinsey.com/~/media/mckinsey/dotcom/client_serv...
[3]https://www.eia.gov/consumption/commercial/pba/health-care.p...
[4]https://electricityplans.com/how-much-electricity-does-an-ap...
[5]https://pmc.ncbi.nlm.nih.gov/articles/PMC5058574/ for water usage per square meter for hospitals
[6]https://pubs.usgs.gov/of/2017/1131/ofr20171131.pdf for water per person per year
The flood of anecdotal thumping seems to distract and undermine deliberation. The sad lack of peer reviewed data and impact analysis hinders thoughtful policymaking.
"More improper redaction shows that the data center using the most water annually is Fireball Group LLC, the Google Data Center in Papillion. Fireball LLC reports 547.88 megagallons for their 2025 Annual Water Consumption."
Edit: The average household in the U.S. uses 10500kWh of power per year. 28.76kWh per day.
https://www.eia.gov/energyexplained/use-of-energy/electricit...
1.2kw is power, not energy (which would be kWh).
The unit you want is spelled “kW”. There really is a difference, especially when you’re inconsistently correcting someone else.
It doesn't have to be really bright lighting but that's the easiest thing to satisfy the underwriter's request, and as a result, the cheapest to implement.
I would have traded a data center for the dairies I lived near as a kid
Maybe we should be calling them "compute centers" instead of "data centers".
> At (edit) 10m, it often exceeds the 90dB threshold for hearing loss.
Why would you be anything like that close if you don't work on site?
A "Rechenzentrum" was a place where lots and lots of people were sitting down crunching numbers with pen on paper. This job also existed in the USA, where originally "computer" was the job title of someone working in such a computer center.
But then electronic computers became better at crunching numbers than human computers, so the job became obsolete. The word stayed around, but the meaning changed. Same way we dial phone numbers despite phones no longer having dials. We also still hang up the phone, despite there being nothing to hang the phone on anymore.
> Why would you be anything like that close if you don't work on site?
I'd also question whether that is inside or outside. We already have regulations that govern the acceptable noise levels around commercial and industrial properties so I highly doubt that a datacenter is going to have an exception.
I do really dislike that AFAIK almost all of these are being powered via burning fossil fuels. Any way you slice it, it's going to mean that locals end up paying higher energy bills.
Sure, millions of people are going to die due to climate change, and that's going to impact the company's profitability, but by then your stock options will have already vested - so that's a problem for someone else!
Personally I will forgive them the Datacenter carbon if they are willing to stop flying private jets.
Well, it was true in the 90s, and it's still true today. Once climate change becomes important enough (that is, causing actual significant damage to important, wealthy countries), it takes just one such country to start a stratospheric aerosol injection project, or, if space launch costs become low enough, put a sunshade between Earth and the Sun. Such solutions have their own risks, of course. But if climate change really becomes bad, they will be tried. And from what I have read, they have a good change of succeeding.
What qualifies as a rich nation anyway? Europe seems pretty rich, and they're melting. One would think the time is now.
They are still not melting enough. Most of the negative consequences of current summer heatwaves can be pretty much solved by installing US-style ubiquitous air conditioning.
An extremely wealthy acquaintance in my local circle is happy to endlessly debate whether climate change is ACTUALLY real or just scientists carving out job security, while flying out for monthly ski holidays wherever there's nice snow in the world when it gets a bit muggy at home.
It’s silly to pretend that technology is anything other than a double edged sword.
I'll grant you social media and advertising, but the other two are not new, and were way worse in the past.
The first problem, in order to reverse a chemical reaction, you need to provide at least the offsetting released energy of the initial reaction. That means for CO2 you need to provide at least as much energy as was released when the carbon oxidized in the first place.
The second problem is that the CO2 is diluted in the atmosphere. Consider the amount of energy needed to remove salt from ocean water, it's not simple.
It just takes a lot of energy to that what was done by burning fossils in the first place.
This is only a problem if AI won't develop cheap carbon neutral energy sources, like fusion or advanced fission, or more efficient solar with cheap grid-scale batteries.
Or to spell it out more clearly, the sarcasm has nothing to do with the truth value of the claim (I’m not making an assertion in either direction).
The rest of us, who might be on green tariffs usually get a commitment that on average almost all our energy comes from renewables. When supply is short then I'm the one who's paying for gas generation rather than the company who has committed to greenwashing. But that also drives investment in renewable generation, in the UK renewables have gone from 2% of power generation in 1991 to around 50% now.
From a climate change perspective, what happens at one particular moment doesn't matter -- it's CO2 emissions globally over time that do.
They are also building solar farms independent of the data centers. People hate it and are against it.
They just built a Costco which turned some cornfields into a giant warehouse. No signs, no protests, no complaints at the city council meetings.
Does that seems unreasonable? We already grow more than enough food in this country and surely we're not getting 300x more value from the corn.
Not everyone is a software developer.
What if the cornfield owners got paid market rates for the power generated when that land is given over to solar? This should be (and ChatGPT agrees) much more profitable given that it is vastly more energy efficient.
Realistically the solution to NIMBYism is to pay the neighbours, but that's not really practical.
NIMBYism isn't rational, it's a manifestation of fear of change, even if they past state of affairs was worse than the future one. I live in a city where the anti-datacentre campaign isn't about some pleasant locale, but a disused toxic industrial land in a steel-mill complex over half a km away from any actual homes.
The corn field is a known quantity.
If we pay each farmer to generate solar (and maybe also stop paying them to grow corn beyond what's actually economically viable), why would they care what the next farm over is doing? I thought we were talking about places where people live too far apart for "backyards".
I understand these companies want to move fast, but this seems to be one of those cases where going slow gets you there faster.
Good luck with this attitude homie!
This was the attitude these leaders took, though. It shouldn’t be downvoted simply for being wrong—it’s wrong in an interesting way.
Lots of folks in Silicon Valley thought they’d be cheered for bypassing bureaucracy. They were wrong and will pay dearly for it, possibly with their wealth and freedom. But it was a commonly-accepted attitude, in part driven by the Bay Area’s obscenely inept bureaucracies around power and building.
"I needed to add a billing address field to the order flow, so I suggested we use Kafka to process shipping requests. My boss said the form field should turn red if an invalid address was entered, and in response, the team lead objected that Oracle charged too much for MySQL. I checked for a pricing page but found a 'call us' form. The CTO decided to go with Kafka, and told us that landlines are an obsolete platform."
Hasn't it been common knowledge for a while -- as in since well before the ai boom -- that those reviews are in practice mostly tools for obstructionism (I don't want anything built) and extortion (fund this project and we'll drop our complaint), and are a significant chunk of why we suck at megaprojects in recent decades?
Is that's even half true, I don't think the "unforced" there really applies.
Shady backroom deals are bad though, and selectively choosing to apply the rules to some and wave them for others isn't cool, in either direction.
If you're only saying we need fewer data centres then your argument is against AI, and not in favor of sensible approach to the climate.
Point being that we've been literally eating our environment to death, long before AI was a twinkle in Scam Altmans eye.
Weird how people never cared then. Weird how vegans were constantly shunned as too extreme then.
In a society of laws you can’t just ignore the ones you don’t like, otherwise it undermines the entire system and suddenly the ones we all generally like/support aren’t being followed either.
You can't just "change laws" when the people building these data centers control the laws.
If the laws aren’t working we should change them. That’s how this all works. Yes we can debate the efficacy but that’s the system we have and just ignoring laws isn’t viable.
The global impact of AI doesn't change when you force DCs to build in another area.
Totally not a new robber baron era at all. That's just fake news guys.
/s
Most of the shorts I've seen about this use an overlayed sound or synthetic sound of some sort.
I've also seen ben jordans price on "infrasound" but that was inconclusive (and not very scientifically rigorous).
I did see one video that measured 60db from the front yard, which I can admit could be problematic, but that fails to show that it's a pervasive or unsolvable problem
https://www.reddit.com/r/interesting/comments/1uleoa2/sound_...
Sound meter shows "only" ~60 db, but that constant audible hum is, IMO, the data center actually assaulting those residents and apparently doing so 24 hours a day 7 days a week.
If someone who lived near that place set fire to the data center and I ended up on the jury and heard these recordings I would do everything in my power to nullify because that seems like a completely justified response to me in the face of the company running the data center not shutting down until the problem is fully fixed.
I think there's more to the story than only AI causing higher electricity bills
2. Most heavy industrial users of power are running stable businesses that are often expected to offset some of the infrastructure costs themselves, making payments over time. Not only have many hyperscalers bent the rules to avoid doing this, even if they did do things the right way, much of the rapid investment is highly leveraged and if a bubble pops, taxpayers would be stuck with the infrastructure bill anyway.
Nobody talks about the local taxes they pay. Its the reason these town councils rush approvals through. My town would chain children to trees to stop a DC, but if they found out it would halve our property taxes, people would volunteer to cut down the trees themselves.
The problem is though that the town council generally wants to do dumb shit pet hobby projects rather than give everyone a tax break.
We hear about their tax breaks though. Why don’t they get treated the same as other local businesses, and pay the same rates?
There is debate about that - and wrapping it all in NDAs makes it hard to assess.
https://www.warren.senate.gov/newsroom/press-releases/ai-dat...
They force these NDAs so people can't collaborate and see what a horrible deal they are getting.
At the same time, these extractive data centers are being used as the primary input to a technology that its most significant leaders say has a substantial chance of killing all humans in the near/mid term, but oh yeah even if it doesn't do that it will wipe out the majority of white collar jobs.
So yeah, color me shocked that most people don't want to see data centers built in their communities. And labeling it "hysteria" is just a lame attempt to discredit real, important, and proven concerns.
More generally, societal fatigue with tech culture overtaking everything, subsuming power from every other corner. Surveillance, the Zuck and Elon, COVID-19, immorality, Luigi Mangione and Elizabeth Holmes, a glut of cat videos and jackal-eared Wikipe-tans, delinquency of minors, 5G and everything else that technology can share some blame for, now focused on their visible manifestations.
The trouble is, once a danger becomes visible and ubiquitous, you’ve already lost. Fighting “A.I. Data Centres” is literally quixotic. If you look up “quixotic” in the Spanish Wikipedia, there is a photo of a chivalrous knight jousting with Nvidia’s CEO.
From what we see on the TV is not that people in rural areas in the US are concerned with the environment. It’s just that they fear energy will become more expensive.
As a non American, I find this incomprehensible.
This is right up there with being able to predict a competitor's total manufacturing capacity for the next year because you have hard numbers on their raw material purchase.
A competitor with that information might have made a rush on a supply-constrained component market or negotiated longer term pricing agreements.
So, given that you don't know what makes your company better then the competition, it's better to hide everything you can so you don't accidentally disclose useful information. God forbid competition discover your own strategic advantage before you do.
Power and water consumption number revelation goes directly to ceiling on maximum compute and relative compute efficiency. It's handing to every competitor on the planet (including China) "Hey, here's the raw situation on how far along we are on our total capacity to do anything."
... whether this should be a trade secret (since power and water are both shared resources with the surrounding community) is a reasonable question, but in the ecosystem where it is, Google having this data public and their competitors not is a problem.
Companies put signs on buildings so you can find them. Signs are important for retailers or offices where people need to find it. This isn't really the case for a data center. Data centers typically have a sign with simply the address of the building on it, so delivery drivers can find it. Any sort of a sign with branding is a waste of money at best, and a security liability at worst.
Even if you aren't worried about intentional security threats to your data center, you don't want the name of your company on it, because people will start showing up at the gate there instead of your primary business location.
At a large enough scale, it doesn't matter how efficient you are - you're still causing a more damage than the region or the world can take.
There being a grain of truth or a valid angle doesn't wash bullshit from being bullshit.
Or here, in this comment from below: https://news.ycombinator.com/item?id=49958135
Now where's the data for the popular claim (which as you conceded is not what you just posted, where it's specifically the overall burden that is too big)? I'm amenable even for a complex mix of cases, where some of the local impact is too much. Would be good to know what share of it is true, even if only a sliver. Although this wouldn't change the bullshit conclusion.
I get it, a lot of the older data centers consumed relatively little power. More than nothing but manageable like 1 or 2MW.
The reason for the "craze" is that the AI datacenters being built are monstrosities that consume way more power, an order of magnitude or more. And a lot of them are going up.
It is a monthly consumption of a poorly insulated house with air conditioning running 24/7. Or around 1344 W of constant consumption.
Older DCs relied on much less efficient compute and cooling, and the only reason why their power usage could be less is simply much smaller scale.
With hyperscalers it of course went a bit beyond that due to the sheer number of racks, but 100MW would've been considered huge.
[0]: https://www.northcdatacenters.com/en/northc-datacenters/delf...
[1]: https://www.eurofiber.com/en-nl/data-centers-netherlands/dat...
A single rack in a corporate server room consumes on the order of 10 to 20kW.
1MWh happens when you run a 1MW load for 1 hour. There are 1000 kWh in 1MWh and 1000 Wh in 1 kWh.
A home with 1344W of constant consumption would consume 32kWh per day.
The data center discussed in the article allegedly consumed 53 MW peak and 13 megagallons of water over a year.
I think the largest data centers consume about 1 GW.
https://www.enconnex.com/data-center-energy-consumption-and-...
https://techplustrends.com/1gw-data-center-power-consumption...
The largest AI datacenters. And there are multiple planned that consume beyond 1 GW. In particular, I'm referring to the box elder county data center which I believe is one of the largest planned. [1]
Prior to these AI datacenters, 100MW would have been considered an extremely large amount of power for a single datacenter.
But also, 1GW is a huge amount of power. More than a lot of states use, and not the really small ones like Wyoming. Just for reference, steel mills use 40MW and there's about 8 of them in the US.
1GW is an absurd amount of power to consume.
[1] https://www.tomshardware.com/tech-industry/kevin-o-learys-9-...
Not even close to a true statement.
EIA says that the lowest electricity use by state in 2024 was Vermont, at 5.5B kwh. That’s about 0.6GW, if you assume continuous uniform use. By the time you get to #48 (Hawaii) you’re over the 1GW mark [1].
According to the EIA, US utility-scale generation capacity in 2025 was 1.3 million megawatts [2], so assuming that these new 1GW datacenters were on the public grid, they would each consume about 1/1300th of the 2025 supply. Not nothing, but not nearly the dramatic impact that you’re suggesting.
For perspective, the Grand Coulee Dam has a generating capacity of about 7,000MW. Robert Moses dam in Niagara Falls does about 2,700MW.
[1] https://www.eia.gov/state/seds/data.php?incfile=/state/seds/...
[2] https://www.eia.gov/energyexplained/electricity/electricity-...
If the Box Elder datacenter were a state it'd rank #19 in terms of power consumption (78 TWh).
It is correct to say these massive data centers will represent a significant percentage of US power consumption when built out.
Attacking them for size is a bit like saying that freight trains are bad because they’re heavy.
Efficient how? Efficiency is a meaningless term without a metric attached. A train is efficient in terms of CO2 used per kg transported, it's not efficient in terms of time to transport.
> Attacking them for size
I didn't attack them for size, I attacked them for energy consumption. And that has multiple knock on effects. From climate change (since a lot of these data centers are being powered by natural gas), to local air pollution, to increased local and regional energy prices. If you are consuming 9GW of natural gas that's 9GW worth of physical gas that's been consumed. Increased demand == higher prices for anything that uses natural gas, heating and power generation come to mind.
I don't care so much if a data center is huge, what I care about is the resources it consumes to operate. Similar to how I don't care how time efficient a private jet is for the wealthy to fly on. I care more about of CO2 inefficient private jets are.
Just some napkin maths. The best combined cycle gas powerplant (Nat gas + steam) have an efficiency around 60 %. So 9 GW of electricity requires enough natural gas for 15 GW (9/0.6 = 15). Natural gas is about 10 kWh/m^3, so in 1 hour, we need 15E9/10E3 = 15E6 m^3 of natural gas. In BTU this is 5.118E10 BTU. A year has 8760 hours, 4.483E14 BTU or 4.483E8 MMBTU in a year.
According to https://www.eia.gov/todayinenergy/detail.php?id=56380&src=%E... the average Nat gas residential consumption in USA is around 80 MMBTU. So these 9 GW of electricity for datacenter s represent the average Nat gas consumption of 5.6 million home in USA.
Meanwhile from a local perspective the fact that a very large datacentre next door is hypothetically aggregating what would have been several other data centres, mostly built elsewhere, is not so reassuring. (Though the other side of that coin is that from the builders' point of view it may be politically easier to get a given amount of capacity built as a few very large plants in places which are less likely or less able to object: regulatory efficiency, distinct from energy efficiency.)
And also conversely, if the mega-DCs (deca-giga-DCs?) really are markedly more efficient, and if they do aggregate lots of new capacity, most especially if they do also end up aggregating existing capacity, that will worsen problems of competition and market concentration in the cloud computing market, at least once the money spigot is turned off again and it's no longer the case that any idiot can get financing for a new datacentre.
If there is demand for it, it's "socially justified". We don't live in a command economy, and thankfully, there's not a panel of woke elders deciding what we get to build. One of the great things about capitalism (indeed, the entire purpose of the thing) is that it does a great job of balancing supply to demand.
The decision of whether or not to build one big datacenter, or a billion teeny datacenters is also well solved by competition, and doesn't need our help.
I haven't done the math but if I see how many tokens I can burn per month for 100 USD it's pretty mind boggling to imagine how much capacity is being added by this build out.
This is not the place to discuss it but even in my wildest predictions I find it hard to imagine that we're really going to need all this capacity anytime soon (although I guess anything spare can go into training).
https://www.pv-magazine.com/2026/09/03/chinas-solar-capacity...
For those who haven't clicked through, that article mentions a 9GW datacentre planned for Utah by Kevin O'Leary's company as well as a 10GW datacentre planned for Ohio by Softbank https://apnews.com/article/ai-data-center-ohio-uranium-enric... and Facebook's planned 7GW data centre in Louisiana https://www.tomshardware.com/tech-industry/artificial-intell... , all largely to be gas-powered it seems.
1 GW is about a single nuclear rector of the Zaporizhzhia powerplant in Ukraine. Other nuclear powerplants in Europe have about 600 to 900 MW per reactor. The newest generation have reactor of 1.4 GW. Just to give an order of magnitude. Every time a 1GW data center is built, to have a continuous supply and CO2 free, it would require to build one nuclear reactor. I find this insane.
Given the current power consumption of states, I find it hard to believe.
Even for smaller sites, 1-2 MW would have been very small.
You see it less now since the shift to the cloud, but in the 00s this stuff was everywhere.
The concern is real and very much deserves the opprobrium voters are showing.
It'd be pretty easy for the company to dispel those outlandish accusations - the PR department could just report the true numbers.
Why not get people riled up over water under the bridge and not unpaid rebates? What can anybody do about it? Send the sheriff to lock out the facility?
https://sustainability.google/files/1699a2e19277e478b38a659f...
The data tables start on p. 90.
Of course, they'd like the water thing to stay an issue as it distracts from others. The water problem has been concentrated on because even some of the most adherent climate change deniers recognise that water access is a significant problem in many areas, so it is an issue that can be usefully raised and won't get tied up in the human-caused-climate-change argument. Unfortunately concentrating on that could be distracting from other matters.
Water use has locally been an issue in places where municipalities were taken advantage of and their local water supply management laws weren't set up to account for a giant consumer to just roll up and start pulling past replenishment rate. But there's nothing about datacenters that demand that be the case; those places were cheap to build in (and why they were cheap to build in should probably be its own graduate political theory thesis / series of fraud and mismanagement investigations and prosecutions).
Both things happened when Sam Altman blogged about how much water the median ChatGPT query used.
Effectively no reward at any risk level, regardless of whether one agrees if there is actually any risk.
If they are using over a certain percent of any public utility (and yes ALL utilities are public) the utility should force it to be disclosed.
Correct. But they can't so easily without Google's (or other big tech companies) big legal departments to make challenging the excuses difficult.
They carefully hid things so they could expand for as long as possible, knowing that having the necessary footprint for services, cloud, and AI would be key to winning the race. When I worked at Google, our leaders were always "proud" they had built such a huge fleet (originally for search + ads but rapidly growing to a wide range of products and research) without competitors knowing just how large, because they figured eventually AWS and MSFT would realize they had to build out massively to handle the coming loads.
So they hide these things to stave off complaints for as long as possible.
I think the disinformation is people thinking that data centers directly use vast quantities of water. What happens is the power use stresses water supplies in local areas.
https://www.theguardian.com/us-news/2026/jun/08/datacenter-a...
Now that it has been, though…
Most of the consumer water is evaporated but that leaves behind dissolved solids and minerals and so some of the water is used to clean that out. That water also often contains additives like biocides, anti-scale agents, and corrosion inhibitors.
This water is discharged into the sewer system or to municipal wastewater treatment facilities.
Here's an article on the various ways data centers can be cooled and what effect each has on water usage [1].
[1] https://www.fwpcoa.org/content.aspx?page_id=5&club_id=859275...
Are they drawing from a flowing source or an aquifer? Is it depleting a limited resource?
Is it evaporative or does the waste get discharged?
Is the waste water treated for the metal that it leaches from the cooling system?
Is it allowed to reach thermal equilibrium first or is the water discharged warm which can cause algae blooms or other problems?
They easily could build near heating system or cooling the water down in another probably big facility close by and send it to the city
While this water usage might be not a big deal in places with abundant water, it could be problematic in some areas of the US.
Nope. While it's technically possible via natural gas fuel cells which are near 100% efficient in combined cycle electric / heat recovery use as compared to our city's existing ~35% efficient combined cycle natural gas fired steam turbine. And one such installation was proposed for this city in the 24MW range. Our beloved local citizens believed it, without reason and in open opposition to openly shared information to the contrary, to be a water guzzling/polluting, pollutant belching, energy consuming, mass surveillance enabling dystopian thing.
Flushed $120M in local investment down the toilet and signed up for worse air quality from ageing inefficient infrastructure because a datacenter is a big faceless thing that's safe to hate.
Right, isn't that what sharing the information is about to show, that it isn't consumed? Show how much water they're adding to the ecosystem, since I'm guessing they'd have numbers on that, if that's how it works.
but trade secrets
I think all data centers and other industrial concerns above a certain size should have to publish that.
The hysteria always gives datacenter usage numbers in a vacuum, never a comparison, for good reason; we might have to start protesting golf courses and paper mills instead.
Traditional data centers have far smaller power density per rack and thus cooling requirements than AI datacenters; if you can use closed loop system without evaporation, water usage is very small. Difference is around order of magnitude
But if you want to fit maximum amount of compute in the space and need to dissipate it, evaporative based system will be far more effective space wise; especially if also for some reason you built it in hot place, rather than somewhere in the north
While on the surface, the proposals sound very efficient and low impact, as always the devil is in the details.
Standard operation would use closed loop cooling, relying on rooftop rain water to keep the loops topped up. However, the company has acknowledged that when ambient temperatures rise above a certain level (26C), they would need to switch to evaporative cooling.
The contract with the government also has a section where the company requires a connection to the state agricultural water supply as a backup [1].
When rainfall is plentiful and temperatures are low, as they often are in tasmania, everything is rosy.
However the state does experience periods of draught and thats where things get sketchy. Conflicts are likely to arise when a large investor in the states economy wants to start draining the local water supply to keep ai agents running at a time when crops and livestock are competing for the same resources.
Similar issues would arise regarding power. Firmus is proposing to underwrite an expansion of the states already extensive Hydro power network, but for that investment they would very likely expect priority access to the grid when supplies start running low.
[1] https://www.tasmaniancountry.com/newspaper/tasmanian-country...
I'm sorry, I never worked in a data center nor lived near one, what's the obvious reason you couldn't dispute them? My only guess would be that Google really wanted to keep the amount of usage secret, but that'd only make sense if it was higher than expected not lower.
Also, they didn't want you to be able to guesstimate how many machines / computing power they had, for similar reasons.
People complain about new data center slurping up all local water and then neither Google nor representatives can say "No, actually we're not, here is the numbers", because Google doesn't want their competitors to guess how many machines they have in one of their data center? I'm sorry but that sounds very far-fetched, must be another reason parent wrote it was "obvious" they couldn't divulge how little/much water the place uses.
It's probably more that they don't want to tell competitors anything they don't absolutely have to.
They published a lot of less specific info (PUE numbers, water efficiency, etc)
They were very sensitive about any site specific numbers.
It is also interesting how much the anti-datacenter panic sounds like astroturf.
Both sides of this issue are doing the same thing, you’d be a fool to believe otherwise.
Creator outreach guy at ControlAI talking about the same type of thing https://lnkd.in/p/ehkuYFkw
Future of life has a program for digital media outreach https://futureoflife.org/project/digital-media-accelerator/
The PACs are well reported on
https://www.axios.com/2026/06/30/ai-safeguards-public-first-...
https://www.axios.com/2026/07/17/ai-safeguards-millions-repu...
Not necessarily astroturfed but definitely biased.
It's not well-paid enough to be immune to kool-aid etc, and it's not on the sociopathic management level either, so we can expect a certain degree of very human rationalization going on.
It offers the compelling counter-narrative of "yes, we're evil, but hell yeah this is lucrative af". That makes the ego even happier, as in that story, one is the clever mastermind playing the system, so we can reasonably assume that that path will be taken.
If that path is unavailable, and cognitive dissonance sets in, the other path left is actually drinking the kool-aid. The corp must be nice, after all.
(This all assumes that people do in fact want to do good and have a conscience, but for those that doesn't apply to, there's the sociopathy path as mentioned in the previous comment. Reality is slightly more murky and dynamic, but I'd argue that this rough if-ladder exists as described.)
That makes it very clear he was an employee of Google working at the data center.
Written communication is hard
All this whinging about "is this project using too much power or not" is pointless scarcity that we've chosen to partake in. If we are to create and spend political capital via hyper-focussing on a single issue, perhaps we should use it to fix this instead?
[1]: https://reneweconomy.com.au/just-staggering-china-installs-1...
This is overstating things.
Even if you only draw power during these three hours of the day, you still need to pay a daily supply charge, which makes up about 30%(and growing) of the average household electricity bill.
You would need about a 15 kWh battery to cover your household usage outside of these three hours.
A 13.5kWh Tesla battery system will cost about $15k installed, the taxpayer will cover about 30% of the battery cost(exc installation), which brings it down closer to $11k.
If you live an an owner occupied detached house(about 50% of Australian households), and you can afford a ~$11,000 for a battery, then you can probably pay off the battery within about 10 years.
[1]: https://thehill.com/homenews/senate/6121776-gop-worried-data...
hint: An individual 18-hole course uses between 300,000 to 1 million gallons per day
If we can close 1 in 3 golf courses and have all the water we need for data centers - and then some - would that be acceptable to our society?
Do we desire so much golf that we'd prefer to stress our water systems instead?
This is a choice, alternative approaches do exist but we should definitely not assume that datacenter water usage is _always_ benign.
You understand, the water isn't used up? It's not even used up in agriculture. Even when you drink a glass of water, the water doesn't actually disappear, soon someone else will be drinking it. The same water that was on the earth when the dinosaurs walked the planet is still on the earth today. No need to panic about someone using water.
The only reason to be concerned is if someone is mixing the water with other chemicals at a rate faster than the natural filtration and then not paying for the treatment to purify the water again. But that is not an issue with data centers. Merely using water is a concern only for people who don't understand how matter is preserved, but a school physics book should dispel those concerns.
> Most data centers are closed loop systems, they circulate the water over and over again.
The linked article talks about >500 million gallons used by one large Google AI datacenter in a year. It's nice that most are closed loop systems, but if it's such a common thing to do then why isn't it done by default with all those AI datacenters? How cheap is water for Google that recirculating it is not a no-brainer?
However, the source of the water can be "used up" faster than it can be replenished, such as in the case of some aquifers
Nope. Plants turn water into oxygen and (the H part of) hydrocarbons, and animals turn those back into water. And after that, the result is different water than the cycle started with.
The current data center craze just feels like the same thing but on a bigger scale.