I lived in a smaller rural town near a Google datacenter. During my career there I had routinely come across some outlandish accusations made about our water/power usage from the locals. But for obvious reasons I could never dispute them. I remember we hired a local who'd grown up in the town and he was "really surprised" when he started to see what we were doing to remain efficient.
The current data center craze just feels like the same thing but on a bigger scale.
Also the unforced error around NDAs and attempts at bypassing environmental and other reviews. People don’t like the idea of their local elected officials bypassing the rules in the dark.
I understand these companies want to move fast, but this seems to be one of those cases where going slow gets you there faster.
I live in northern va and I can tell you that histeria is about ballooning electric and other bills. vast majority of population is unfortunately “not there” in regards to any unrest about AI
My biggest problem with the data centers getting built around me are that they are building new natural gas power plants to power them, that will be greenhouse gases and decrease local air quality. I would be a lot more ok with it if they were using wind or solar. And maybe built it somewhere that didn't have a water shortage.
Google considers pretty much everything in hwops to be a trade secret. The point is to keep the specifics unknown as much as possible from competitors; the public at large's knowledge is just collateral damage in that fight.
And this one in particular is a high-ticket trade secret; it has been for decades, long before the AI push.
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.
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.
> I think the largest data centers consume about 1 GW.
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.
> But also, 1GW is a huge amount of power. More than a lot of states use
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.
> 1GW of power datacenters, more efficient GPUs, more efficient training, Xiaomi/Deepseek disk based caching + other efficiency gains.
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).
> But for obvious reasons I could never dispute them.
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.
Google never wanted to share details about their datacenters because they thought (think?) their power efficiency was a competitive advantage. So it'd be lower than you expect.
Also, they didn't want you to be able to guesstimate how many machines / computing power they had, for similar reasons.
Right, they're so power efficient they don't want to tell the people around them how little water they consume or "not consume"?
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.
Tbh this sounds perfectly normal to me. Companies often don't think it's worth defending themselves if theres no actual financial damage, and certainly don't want random employees to be doing it. It can be very frustrating, both for employees who want to defend their reputation and for the public who want the truth, but companies can be extremely conservative about chosing their public relations battles.
Along those lines just a reminder folks, China is installing about 1GW equivalent of Solar power per day [1]. Lot of Australians can simply stop paying for power now due to free power on offer and batteries [2]. These are not zero-regulation, socialist wastelands.
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?
The same one that is paying for their support for data centers and is shifting as a response [1]. Like I said, if political capital is being spent shifting peoples' stances, how much of a waste is it to know we could move them on this and we chose not to. Generational fumble.
Most data centers are closed loop systems, they circulate the water over and over again.
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.
Sure the water is still somewhere, but its location is important for future use. US water reservoirs are at problematically low levels already.
> 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?
That’s not entirely correct. Overall water is preserved but local sources of freshwater can be depleted or contaminated, which raises the cost of water.
It is worth remembering that it heavily depends on cooling system involved.
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
I've been spending some time in Tasmania, Aus for the last few weeks and have been following the local debate regarding 3 data centres Firmus is planning on building here.
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.
Are you sure? People think water for data centers is "consumed", like the output is sewage or something. Giving those people concrete numbers is not gonna change such misconceptions (and that's assuming they look at such numbers in reasonable relative terms in the first place).
The majority of current large data centers use open-loop evaporative water cooling which does in fact consume water.
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].
Seems easy enough to report multiple numbers and be clear on what water comes in (can even separate initial filling/construction from ongoing usage) vs what gets lost and what comes out as waste. I think people can understand this and it isn't a good enough reason to not share the numbers.
Evaporated water stays in the water cycle though. It comes back down as rain, and returns to the water sources that were depleted (assuming these data centers are generally not using underground aquifers, which I think is true).
Fair point. But if I had initially learned that google was putting data centers in and it recycled water and it was a minor amount I would be more receptive than hearing that google is bullying communities and hiding details behind NDAs.
> Are you sure? People think water for data centers is "consumed", like the output is sewage or something.
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.
Not sure why this is being downvoted. If you noticeably distort the local economy then people are naturally going to want to know what resources you're consuming.
but trade secrets
I think all data centers and other industrial concerns above a certain size should have to publish that.
The irony is that it is unlikely datacenters would hit that "above a certain size"
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.
Surely the people who have worked in tech, with datacenters or even in them for decades are wrong, and the people who learned the word "datacenter" 6 months ago are right.
> I had routinely come across some outlandish accusations made about our water/power usage from the locals. But for obvious reasons I could never dispute them.
It'd be pretty easy for the company to dispel those outlandish accusations - the PR department could just report the true numbers.
Worth considering in the current climate. My impression a while ago is that the exact water and power usage of any large company's data center is seen as a trade secret of sorts; if a large company invested to become XX% efficient, it might not want competitors to know it hit a certain efficiency target.
Largely because of how many times it could have been stated prior, given it had been a talking point for some time, and the lack of actual evidence (him saying something is not often taken at face value…).
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.
Data centers tend to be higher security facilities for a reason. Publicizing any of their technical specs isn’t really beneficial, especially when any numbers that are provided will be dismissed unless even more information is provided to back them up.
Effectively no reward at any risk level, regardless of whether one agrees if there is actually any risk.
The big hyperscalers always knew that eventually the tide would turn against them (for example when local towns noticed that the DCs didn't actually need very many locals to run the facilities).
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.
It’s not high security if everyone is mandated to do it. Power consumption, energy sources, water usage for coolant, waste water treatment plan, etc.
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.
In that specific position probably leaning more towards biased than towards well-informed, though both components likely do exist.
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.
Being paid well enough gives the mind an out from cognitive dissonance that doesn't involve actually believing the company narrative.
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.)
So, basically, not a meaningful amount of water at all. Credit to the story authors for taking the time to frame what 13MM gallons actually works out to.
Yes. As another comparison point, 13 million gallons per year amounts to 24 gallons per minute which is about 5 (five) bathtub faucets running 24/7/365.
The water bill alone would be only $100k/yr assuming average US domestic rate of $8 per 1000 gallons.
$87k household. So that's 6 median incomes of water being used up by that datacenter. Put that way it doesn't sound bad, especially when you glance at the electric bill.
Also when it evaporates instead of going to the sewer that drops the price in half.
Edit: Oops I got that mixed up and was treating it like a per-faucet cost. So actually it's one full household income converted into water to handle the entire facility.
Do you think people look at big facilities and calculate that they're spending whatever times the medium wage on a crucial service and get mad about it?
Converting things into median wages is a way to put them into scale. They spent this many wages for construction, they spend this many on supplies, they sell this many in product, they hire this many people to work there. If anyone thinks its bad for anything a company buys to exceed a median wage, that's a very silly expectation and I wonder where they got it.
Perhaps not a meaningful amount at a national/worldwide scale. However, an individual datacentre consuming that amount at a local/regional scale could be more impactful.
I don't understand your question. I'm commenting on the figures we know from this report. I don't have to work out the motives behind them to evaluate whether they're material (they are not).
Please elucidate if I'm wrong, but, isn't the problem the distribution of the water and not the amount? If you install a building of people in a road that is not prepared for that, all people down the people are in risk of lack of water.
I think at some point someone used the argument of the spoon of water to say about the quantity, but I remember to see explanations that it was unrelated.
> If you install a building of people in a road that is not prepared for that, all people down the people are in risk of lack of water.
Part of why construction of such buildings (starting with pursuing and getting a permit for it) takes as long as it does is because those issues are worked out in advance.
Nobody who is responsible for managing water is surprised by data centers appearing, it's not like they just grow overnight after a rain like people believe mushrooms do.
This feels like an important point! People comparing to leaks (or other distributed examples that aggregate similarly) are not thinking about how the amount of water used in a single location changes the scale.
It looks to me like there's two Google DCs with different LLCs?
> However, using a computer cursor to highlight the redacted text box in the report to DWEE, then copying and pasting it into a separate document reveals that *Agate LLC* is reporting 52.65 megawatts of electricity during peak electrical demand and 13.299 megagallons of water between cooling towers, evaporative systems and site operations in the last year. That totals out to 13 million gallons of water.
> For context, 13 million gallons is enough to fill around 20 Olympic-size swimming pools, and is less than half of what the City of Lincoln reports using on Sept. 29.
> The *six* data centers reporting as of Sept. 30 show a total use of 765 million gallons of water last year, enough water to fill 1,159.93 Olympic-size swimming pools.
> 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.
Even that 765 is like 0.04% of the water needed to irrigate Nebraska's ethanol corn crop alone. And ethanol is basically an agricultural proof-of-work function.
No, it’s that if there’s nothing to hide, then they shouldn’t jump through so many hoops to avoid basic industrial reporting requirements.
It’s not any one thing, it is the litany of things they do in order to avoid any transparency or accountability. It is undemocratic and contradictory to the idea of good governance of public resources.
If anything, a little sunshine here would probably be the best medicine. Trying to be opaque is what allows conspiracy theories to exist in the first place.
I'm commenting here about the water figure. Are you talking about something else? Because the impact of that amount of water usage is knowable regardless of the corporate structure of the data center.
Because nobody reasonable believes water use is any issue, but these numbers are proxy for the usual stuff companies don't disclose for usual reasons (ones that rhyme with "competition" and "stock market")?
They sure as hell aren't making it easy for the "reasonable" to take their side when their response to the "unreasonable" is "the numbers you claim are bad are actually good, but you don't get to see them because they go to a different school".
Because hedge funds are going to use those numbers to get better estimates of their business and revenues and use that in their forecasting which means Google loses in the stock market.
Are you regularly filling Olympic sized pools and checking the water usage of an entire town in Nebraska? I would rather see this put into context by comparing it to other businesses normalized by power usage or square footage.
I can see how a 0.27% increase in Lincoln’s water usage intuitively seems insignificant, but that was a a low usage data center in a large city. Papillion has 1/12 of Lincoln’s population, yet it’s data center used 42x more water. Assuming water usage was proportional to population it would have increased the city’s water usage ~136%.
It doesn't seem insignificant. It is in fact insignificant. And it supports one of the most popular and widely used consumer products of the last 100 years, so it seems like a pretty incredible bargain.
Tech companies are bundling AI into their base subscriptions and force feeding it to consumers, so I doubt the metrics they are reporting to their shareholders represent real demand or value.
Yes. GP is literally saying that 1-2 million gallons of water is a pretty small number not meaningful to anything at scale.
It's basically like framing power consumption in watt hours. A million watt hours is tiny at any scale, your whole neighborhood uses millions every day.
Please compare 1 million gallons to the 1 quadrillion gallons in the aquifer, or the 31 trillion that falls from sky each year in Nebraska and get back to me.
Nebraska's main product is corn. Irrigated corn takes roughly 2 acre-feet of water, so the total DC water usage in the article (765M gallon) could irrigate 1k acres of corn. That's a single average-sized farm, not even a big operation. Calling a rounding error might be too generous.
People shouldn't be latching onto side effect costs of AI data centers like water or energy use, it seems like an abstraction layer above the core complaint. If we don't want AI, or LLMs, or data centers that should be the complaint. Once it is abstracted to water use or similar it opens the door for data to show the actual water use is comparatively small.
Its like arguing about climate change by zeroing in on carbon emissions. The issue people should be raising is wasted energy, over consumption, not knowing "enough", etc. Once the issue is pinned to a single externalized factor the argument lives or dies by that factor.
This article seems to focus on one of the less interesting in terms of water usage data centers (for most of us non-locals) because the paper is from Lincoln and that data center is in Lincoln.
The article's data center used 13 mil gal of water, but as they point out another one used up more than 500 mil gallon.
Apparently Google self-reported the total 700+ million, so the main thing to learn from redacted numbers is how much smaller all the other data centers are compared to the 500 mil gallon one.
Update: in terms of water usage. Not sure whether 500 mil gallon one is larger or just really water-inefficient. That would be a nice thing for some journalist to research...
Update 2: it seems that Google uses water-evaporation cooling in regions that they don't consider water-stressed. Here's a not-so-great reference (scroll down if the text-fragment link doesn't work): https://www.axios.com/2026/06/03/google-pushes-water-standar...
The concern around data centers looks a lot like a moral panic. They've been around for decades and nobody cared before the backlash against AI. They're like any other industrial building, subject to the same resource constraints and zoning laws. I suspect if the owners of that land built a steel mill, a semiconductor fab, or anything other than a data center with similar power and water requirements, nobody would care.
You are overlooking scale and pace. No one is trying to build more steel mills in the next 3 or 4 years than have been built over the entire history of the steel mill industry so far, with each new steel mill using around an order of magnitude more energy in the same space than steel mills built a few years ago.
I'm not against local governments setting zoning laws and regulations if they're concerned about the pace of growth. Personally, I'm not concerned because I think we can and should invest in infrastructure to keep up with demand. This will benefit all industries, not just AI. My point is that I suspect (I'll admit that I haven't done any research on this) that people wouldn't be against facilities with the exact same footprint if they didn't carry the label "data center" and this is primarily driven by a backlash against AI rather than the resource use itself. In that sense, they're trying to rebrand a political/culture war issue as a zoning or utility issue.
There is sometimes incorrect conflation between what the datacenter is permitted for and what the datacenter actually draws.
> Many articles about current or future AI data centers report the number in the water permit they apply for as the amount of water they actually use day to day. But this almost never happens. [1]
This is from the problematic and often wrong but in this case correct Andy Masley. It's not completely clear to me whether this number is observed usage or a theoretical permitted maximum, in which case the actual usage could be even lower. It could be either.
He's done just terrible things, the accusations are true. No one wants to associate with him because of all the bad things he's done. Everyone knows he's a liar, and anyone asking for evidence of it is a perfidious pro-AI ideologue.
He’s a Pro-AI ideologue who engages in the same motivated reasoning he complains about others using. His piece on heat island effects fell over under scrutiny because he just doesn’t have the analytical background to actually do these investigations.
This is entirely dependent on where you are geographically in terms of rainfall and weather. Both water and energy are abundant. The trick is getting them where you need them. So the where is what matters.
Seems like it'd depend a lot on e.g. carbon impact of the marginal watt v.s. environmental impact of the marginal gallon? Might depend on the individual site.
Is there much impact of a marginal gallon? This is evaporative cooling, no? It's gonna come back down as rain and refill wherever datacenter took it from.
Depends on where it's coming from and who else needs water. If the water supply being drawn from was collected from rainwater initially and there's still enough for local residents to also have water, then sure, it's probably sustainable. But if the water is coming from an underground aquifer and there's otherwise not enough rain in the region to go around then it's a finite resource.
I dont like the olympic swimming pool comparisons so I converted to immigrant population comparisons. How many people's worth of electricity and water are the data centers consuming , of course all the analysis was done on gemini since we are talking about google data centers :) here's the output
> Data centers impact local utilities unevenly: while their water consumption is roughly equivalent to adding a small town of 25,000 people, their electrical footprint across the state exceeds the residential power needs of the entire population of Nebraska.
> Together, Google’s three Nebraska data center sites draw enough electricity to power up to 290,000 homes—equivalent to supplying the entire residential power grid for roughly one-third of Nebraska’s total population.
> Agate LLC reports expecting a refund of $55,822,472 from 2025 taxes. Fireball LLC is expecting a refund of $39,171,573.39 from 2025 taxes. Westwood Solutions LLC, the Google-owned data center in Omaha, is expecting a refund of $22,558,881 from 2025 taxes.
Regardless of the comments below there seems to be a crazy ideology forming about being completely pro or anti AI. On both sides people are either under or ober reporting how much power these datacenters actually use
kind of confusing numbers here. the 13 million gallon figure refers to one specific datacenter, but you add on five more datacenters and the number grows to... 765 million gallons? The datacenter in Papillion accounts for 548 million of those. Is it bigger? Is it less efficient? not a lot of info apparent in this report.
If those numbers are yearly, they are significant (about what the city of Lincoln uses in a month) but perhaps not troubling. Depends on your perspective.
Roughly 40 acre feet of water for the Lincoln DC. An acre of corn uses maybe a couple of acre feet a season. Corn irrigation alone uses tens of millions of acre feet of water per season in the USA.
I don't really understand the hand-wringing about data centers and water. Unlike fossil fuels, when you "use" water, it doesn't go away, it goes back out into the environment. And since DCs use it for cooling, not some industrial process, it doesn't even contaminate the water in any meaningful way.
The one exception is if they're using fossil water pumped from aquifers, which doesn't renew (back into the aquifer that is), but nobody seems to bother to make the distinction.
True if you're using a closed-loop system. But if you're using evaporative cooling (which is the main cause for concern) then the place where the water 'comes back into the environment' may be quite far away (at least in a useful sense).
I think the bigger issue with water usage ( if it's actually an issue ) is when the system is not able to handle the additional demand ( drying up wells, etc. )
Especially during the summer where water supply is already strained in some areas.
Yes water goes eventually back into circulation but that does not help in the place where the system is already at it's limit
20-25% of the water consumed for evaporative cooling becomes contaminated waste water. It is used to clean out the residue left behind by the 75-80% of the consumed water that is evaporated.
The contaminants consist of dissolved solids and minerals that were in the incoming water, plus additives such as anti-scalers, corrosion inhibiters, and antibacterials.
Mostly, but not entirely. It is a huge issue in the Southwest, and it seems like a lot of DCs are being proposed and/or built in areas where water is far from abundant.
Fresh water supply has been on a downward trend for years. We are already looking at a future without enough fresh water, before the new datacenters come into play. The datacenters are being installed in some of the driest areas of the country that were already water stressed. In some cases they can also contaminate local water systems.
I’m not disagreeing with that but it fundamentally feels like a misdirection. It’s easily hand waved away eg “it uses less than a golf course” and yet it’s still a constant media talking point.
If you pump water out of one watershed into another, perhaps via evaporation and wind transport, then you better expect the downstream users, deprived of their water, to complain.
For example, the upstream users of water from the Colorado use so much that Colorado River Delta is nearly always dry. Or read about the California water wars.
Further, increased competition for surface water raises prices, which encourages additional aquifer extraction.
The problem in California is that competition for water doesn't raise prices. Much of our water supply is allocated based on property rights rather than as a commodity so prices for some users remain artificially lower than the market price. This causes waste and shortages. Eventually we'll have to reform the system but there's no political will to do so.
The water for the datacenter comes from the same place residents, farmers, etc get their water. If there is already not a lot of water, the datacenter can cause running out of water. Sending used water downstream doesn't magically reappear upstream.
> For context, 13 million gallons is enough to fill around 20 Olympic-size swimming pools, and is less than half of what the City of Lincoln reports using on Sept. 29.
Over the course of a year, that doesn't sound terrible.
That's wishful thinking, I know, but would love a breakdown of those metrics into what actually goes into AI inference, training versus just TikTok / YouTube streaming, or just a Google search...To have some perspective. My hypothesis that prolonged video streaming might contribute similarly to AI usage? (But CDN etc might make that tricky to calculate...)
It's not terrible at all. It's dwarfed by common municipal demands, let alone industry or agriculture. It's an amount of water that's basically not worth talking about.
Data centers are ugly and dystopian. We can likely survive as a species without building a single other one. Society and general well-being does worse because of them and what they bring.
From a redaction point of view: I can't believe this is still happening. Surely by now all government and local government offices have a data/records person and policies about information handling. I guess it's not really sensitive info like the name of a government informant. Maybe they intentionally leaked it with a 'whoops, we tried' excuse.
If it really matters, block it out, print it, scan it, then send that.
If it really really really matters, block it out with random extra amounts of width, print it, scan it, edit the page images in paint to move the text to the right of a redaction forward or back a bit so people can't guess the words based on width.
It… doesn’t look that bad. Electricity use is sizable. Water which usually gets all this screaming is, well, look up how much a golf course or a large hotel uses.
It varies pretty greatly, but our golf course uses untreated water from a relatively dirty river that is not used for drinking. It looks like Google is using ground water from the aquifer. These are significantly different beasts.
Not a conspiracy theorist, but wait that really sounds effective. Just few months back I was wondering why were big datacenter companies not trying to sway the public that the water usage is low per query or whatever metric.
There will be protests and cases related to water usage of AI, but none of them would hold any ground. And winning the water case would just make the argument in their favour.
Historically large data centers used evaporative cooling. So the water disappeared in sense that it's no longer available in a useful form in it's original locations. That's changing with growing use of closed-loop systems.
A lot of people think it does, surprisingly. People memed too hard about "querying ChatGPT drains lakes" and anyone who contested it was deemed a booster, so it stuck.
If everything was on the up and up they wouldn’t be hiding things behind shell companies and ndas. The water and electric usage is the least of our worries. These things will make cancer alley look like a Sunday picnic. There is a reason executives won’t live anywhere near the hot zone.
I lived in a smaller rural town near a Google datacenter. During my career there I had routinely come across some outlandish accusations made about our water/power usage from the locals. But for obvious reasons I could never dispute them. I remember we hired a local who'd grown up in the town and he was "really surprised" when he started to see what we were doing to remain efficient.
The current data center craze just feels like the same thing but on a bigger scale.
The current hysteria has little to do with datacenters and a lot to do with societal unrest about AI.
Also the unforced error around NDAs and attempts at bypassing environmental and other reviews. People don’t like the idea of their local elected officials bypassing the rules in the dark.
I understand these companies want to move fast, but this seems to be one of those cases where going slow gets you there faster.
That can't be true. People obviously believe DCs have big local impact and it really does affect their power bills at least.
The global impact of AI doesn't change when you force DCs to build in another area.
The noise issues seem real, but don’t impact whole cities
I live in northern va and I can tell you that histeria is about ballooning electric and other bills. vast majority of population is unfortunately “not there” in regards to any unrest about AI
Yup. From Fairfax. Very much what you said.
My biggest problem with the data centers getting built around me are that they are building new natural gas power plants to power them, that will be greenhouse gases and decrease local air quality. I would be a lot more ok with it if they were using wind or solar. And maybe built it somewhere that didn't have a water shortage.
> The current data center craze just feels like the same thing but on a bigger scale.
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.
Except that's not what you see people be up in arms about, or news reporting. Just like with a lot of other topics.
There being a grain of truth or a valid angle doesn't wash bullshit from being bullshit.
It’s not the water. It’s the electricity usage, air and noise pollution from sanctioned (and often unsanctioned) diesel generators.
The concern is real and very much deserves the opprobrium voters are showing.
Those reasons aren’t obvious to me. Why not just be transparent if that’s the case?
Google considers pretty much everything in hwops to be a trade secret. The point is to keep the specifics unknown as much as possible from competitors; the public at large's knowledge is just collateral damage in that fight.
What benefit could competitors get from that information?
[delayed]
The ability to improve their data centers, of course. Google is the company that launched data centers with commodity components.
And this one in particular is a high-ticket trade secret; it has been for decades, long before the AI push.
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.
If they are so good at it, they should push to make that the bare minimum and get to hold all future datacenter permits.
Yep - better off redacting the great efficiencies we’re doing.
Did your datacenter consume 9GW?
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.
There were certainly 1GW DCs in the US 10+ years ago.
> Did your datacenter consume 9GW?
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...
> I think the largest data centers consume about 1 GW.
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-...
> But also, 1GW is a huge amount of power. More than a lot of states use
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.
[1] https://www.eia.gov/state/seds/data.php?incfile=/state/seds/...
[2] https://www.eia.gov/energyexplained/electricity/electricity-...
Fine, an exaggeration but not off by much.
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.
That ought to be a good thing. Energy is used to do work. We just need to do it responsibly.
> 1GW of power datacenters, more efficient GPUs, more efficient training, Xiaomi/Deepseek disk based caching + other efficiency gains.
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).
It’s not ‘spare goes to training’. These data centres are training first. Inference is really rather efficient, so not why datacenters are so big.
If we printed electricity like China, this would not be an issue, we can have more electricity at lower prices when we're ready
https://www.pv-magazine.com/2026/09/03/chinas-solar-capacity...
> But for obvious reasons I could never dispute them.
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.
Google never wanted to share details about their datacenters because they thought (think?) their power efficiency was a competitive advantage. So it'd be lower than you expect.
Also, they didn't want you to be able to guesstimate how many machines / computing power they had, for similar reasons.
Right, they're so power efficient they don't want to tell the people around them how little water they consume or "not consume"?
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.
I worked at google as well.
They published a lot of less specific info (PUE numbers, water efficiency, etc)
They were very sensitive about any site specific numbers.
Tbh this sounds perfectly normal to me. Companies often don't think it's worth defending themselves if theres no actual financial damage, and certainly don't want random employees to be doing it. It can be very frustrating, both for employees who want to defend their reputation and for the public who want the truth, but companies can be extremely conservative about chosing their public relations battles.
Along those lines just a reminder folks, China is installing about 1GW equivalent of Solar power per day [1]. Lot of Australians can simply stop paying for power now due to free power on offer and batteries [2]. These are not zero-regulation, socialist wastelands.
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...
[2]:https://www.energy.gov.au/rebates/solar-sharer-offer
Hmmm, which political party is famously against solar and renewables in general?
The same one that is paying for their support for data centers and is shifting as a response [1]. Like I said, if political capital is being spent shifting peoples' stances, how much of a waste is it to know we could move them on this and we chose not to. Generational fumble.
[1]: https://thehill.com/homenews/senate/6121776-gop-worried-data...
China are also building tons of coal plants.
What about water usage?
For most datacenters the actual largest source of water use is indirectly from power generation (coal, nat gas). Solar reduces that a lot also.
Funny how nobody asks the folks who build golf courses "what about water usage?"
hint: An individual 18-hole course uses between 300,000 to 1 million gallons per day
I think people actually do though? They are even sabotaged at times (has happened in Europe at least), to protest against their water usage.
I’ve seen it brought up many times
Rather than talking in absolute numbers, ratios help relate human activities better, an example:
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?
The water isn’t actually used. It is slightly warmed.
In some cases chemicals are added to the water to increase cooling efficiency.
This is a choice, alternative approaches do exist but we should definitely not assume that datacenter water usage is _always_ benign.
Most data centers are closed loop systems, they circulate the water over and over again.
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.
Sure the water is still somewhere, but its location is important for future use. US water reservoirs are at problematically low levels already.
> 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?
> It's not even used up in agriculture.
However, the source of the water can be "used up" faster than it can be replenished, such as in the case of some aquifers
That’s not entirely correct. Overall water is preserved but local sources of freshwater can be depleted or contaminated, which raises the cost of water.
It is worth remembering that it heavily depends on cooling system involved.
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
I've been spending some time in Tasmania, Aus for the last few weeks and have been following the local debate regarding 3 data centres Firmus is planning on building here.
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...
Ok and? This wouldn’t be an issue if google et al would just share the numbers instead of hiding behind NDAs and whatever else.
Are you sure? People think water for data centers is "consumed", like the output is sewage or something. Giving those people concrete numbers is not gonna change such misconceptions (and that's assuming they look at such numbers in reasonable relative terms in the first place).
The majority of current large data centers use open-loop evaporative water cooling which does in fact consume water.
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...
Seems easy enough to report multiple numbers and be clear on what water comes in (can even separate initial filling/construction from ongoing usage) vs what gets lost and what comes out as waste. I think people can understand this and it isn't a good enough reason to not share the numbers.
There are so many facets that numbers alone don't cover:
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?
It is consumed because they evaporate it away and remove it from the local environment.
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
Evaporated water stays in the water cycle though. It comes back down as rain, and returns to the water sources that were depleted (assuming these data centers are generally not using underground aquifers, which I think is true).
You can't control the water cycle over a datacenter and its water source, it is a global system.
Fair point. But if I had initially learned that google was putting data centers in and it recycled water and it was a minor amount I would be more receptive than hearing that google is bullying communities and hiding details behind NDAs.
> Are you sure? People think water for data centers is "consumed", like the output is sewage or something.
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.
Not sure why this is being downvoted. If you noticeably distort the local economy then people are naturally going to want to know what resources you're consuming.
but trade secrets
I think all data centers and other industrial concerns above a certain size should have to publish that.
The irony is that it is unlikely datacenters would hit that "above a certain size"
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.
People do just that.
Industrial use should never get to hide. It’s such impactful use so every aspect should be scrutinized.
You have some very motivated people with deep pockets who do not like negative press. Say anything negative and the downvotes will roll in.
Surely the people who have worked in tech, with datacenters or even in them for decades are wrong, and the people who learned the word "datacenter" 6 months ago are right.
Facts and experience mean nothing when everyone knows you are a liar.
> I had routinely come across some outlandish accusations made about our water/power usage from the locals. But for obvious reasons I could never dispute them.
It'd be pretty easy for the company to dispel those outlandish accusations - the PR department could just report the true numbers.
Worth considering in the current climate. My impression a while ago is that the exact water and power usage of any large company's data center is seen as a trade secret of sorts; if a large company invested to become XX% efficient, it might not want competitors to know it hit a certain efficiency target.
They do! In great detail:
https://sustainability.google/files/1699a2e19277e478b38a659f...
The data tables start on p. 90.
Sam Altman has stated the water usage in multiple interviews (it's low) and people don't believe him.
That’s because he’s a known liar. Like he’s famous for his sociopathic lying even amongst sociopathic lying billionaires. Impressive almost.
Largely because of how many times it could have been stated prior, given it had been a talking point for some time, and the lack of actual evidence (him saying something is not often taken at face value…).
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.
That would not be good PR because a) it would be seen as flippant and b) no one would believe Google.
Both things happened when Sam Altman blogged about how much water the median ChatGPT query used.
Data centers tend to be higher security facilities for a reason. Publicizing any of their technical specs isn’t really beneficial, especially when any numbers that are provided will be dismissed unless even more information is provided to back them up.
Effectively no reward at any risk level, regardless of whether one agrees if there is actually any risk.
Absurd! every company ever would love to hide behind these same excuses.
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.
> every company ever would love to hide behind these same excuses.
Correct. But they can't so easily without Google's (or other big tech companies) big legal departments to make challenging the excuses difficult.
How does hiding water usage protect the facility exactly?
The big hyperscalers always knew that eventually the tide would turn against them (for example when local towns noticed that the DCs didn't actually need very many locals to run the facilities).
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.
It’s not high security if everyone is mandated to do it. Power consumption, energy sources, water usage for coolant, waste water treatment plan, etc.
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...
Google has some custom hardware, so directly reporting it could be an issue.
Now that it has been, though…
it's interesting how much this sounds like astroturf.
> it's interesting how much this sounds like astroturf.
It is also interesting how much the anti-datacenter panic sounds like astroturf.
And who exactly would they be astroturfing for?
Big Labor j/k
ignorance
China and Russia.
People looking for something to run against?
LinkedIn CV says 8 Years working for google as a Data Center Facilities Technician, so. Yeah.
Not necessarily astroturfed but definitely biased.
Biased, or well informed?
In that specific position probably leaning more towards biased than towards well-informed, though both components likely do exist.
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.
How would being high paid help make you immune to kool aid? In my experience in low paid work, it’s a lot easier to hate the company…
Being paid well enough gives the mind an out from cognitive dissonance that doesn't involve actually believing the company narrative.
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.)
Have you ever talked to people with a low paying job?
Yes, I find it a lot more pleasant than most of this industry.
Also outright dishonest to not add that as a disclaimer!
>> During my career there
I certainly read that as "there" being the town, not the actual data center!
" I remember we hired a local who'd grown up in the town and he was "really surprised" when he started to see what we were doing to remain efficient"
That makes it very clear he was an employee of Google working at the data center.
Same, but revisiting it with context, it does seem mostly obvious enough.
Written communication is hard
He specifically said he worked there. Did you think perhaps he possibly worked for the outsourced cleaning crew?
The use of the word "we" seems pretty clear to me.
So, basically, not a meaningful amount of water at all. Credit to the story authors for taking the time to frame what 13MM gallons actually works out to.
Yes. As another comparison point, 13 million gallons per year amounts to 24 gallons per minute which is about 5 (five) bathtub faucets running 24/7/365.
The water bill alone would be only $100k/yr assuming average US domestic rate of $8 per 1000 gallons.
So a system to recirculate the same 24 gallons per minute would be more expensive than $100k per year for water?
My state's gasoline refineries use about ~13 million gallons of water each day. It's about .7 gallons of water per gallon of gasoline refined.
I don't know if you mean to be reinforcing that it's nothing or not
Remind me median income?
$87k household. So that's 6 median incomes of water being used up by that datacenter. Put that way it doesn't sound bad, especially when you glance at the electric bill.
Also when it evaporates instead of going to the sewer that drops the price in half.
Edit: Oops I got that mixed up and was treating it like a per-faucet cost. So actually it's one full household income converted into water to handle the entire facility.
Yeah I’m sure someone on median income feels the same
Probably, yeah.
Do you think people look at big facilities and calculate that they're spending whatever times the medium wage on a crucial service and get mad about it?
Converting things into median wages is a way to put them into scale. They spent this many wages for construction, they spend this many on supplies, they sell this many in product, they hire this many people to work there. If anyone thinks its bad for anything a company buys to exceed a median wage, that's a very silly expectation and I wonder where they got it.
Remind me who leaves 8 bath tub faucets running 24/7/365?
Me?
I live in a country of 5 million people and we lose 40 times that a year to leaks.
So a single datacentre introduces a water consumption cost that is within an order-of-magnitude of a near-median-population nation.
Perhaps not a meaningful amount at a national/worldwide scale. However, an individual datacentre consuming that amount at a local/regional scale could be more impactful.
Why aren’t the data centres just honest about their water usage if it’s seemingly so low?
Transparency is not the same thing as honesty.
Companies might want to avoid this being used as a side channel to infer the amount of their compute.
Companies might want to avoid setting a precedent where activists can demand such information.
Companies might want to avoid the effort of publishing this data, or the risk of publishing incorrect data.
I don't understand your question. I'm commenting on the figures we know from this report. I don't have to work out the motives behind them to evaluate whether they're material (they are not).
Because the marketing teams don't do numbers and it's a high risk strategy
They are, and people just accuse them of lying!
Facts & reason do nothing to a moral panic.
Please elucidate if I'm wrong, but, isn't the problem the distribution of the water and not the amount? If you install a building of people in a road that is not prepared for that, all people down the people are in risk of lack of water.
I think at some point someone used the argument of the spoon of water to say about the quantity, but I remember to see explanations that it was unrelated.
> If you install a building of people in a road that is not prepared for that, all people down the people are in risk of lack of water.
Part of why construction of such buildings (starting with pursuing and getting a permit for it) takes as long as it does is because those issues are worked out in advance.
Nobody who is responsible for managing water is surprised by data centers appearing, it's not like they just grow overnight after a rain like people believe mushrooms do.
This feels like an important point! People comparing to leaks (or other distributed examples that aggregate similarly) are not thinking about how the amount of water used in a single location changes the scale.
“Fireball Group LLC, the Google Data Center in Papillion… reports 547.88 megagallons for their 2025 Annual Water Consumption”.
I think the 13M is referring to “during peak times”, whatever that means.
It looks to me like there's two Google DCs with different LLCs?
> However, using a computer cursor to highlight the redacted text box in the report to DWEE, then copying and pasting it into a separate document reveals that *Agate LLC* is reporting 52.65 megawatts of electricity during peak electrical demand and 13.299 megagallons of water between cooling towers, evaporative systems and site operations in the last year. That totals out to 13 million gallons of water.
> For context, 13 million gallons is enough to fill around 20 Olympic-size swimming pools, and is less than half of what the City of Lincoln reports using on Sept. 29.
> The *six* data centers reporting as of Sept. 30 show a total use of 765 million gallons of water last year, enough water to fill 1,159.93 Olympic-size swimming pools.
> 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.
Even that 765 is like 0.04% of the water needed to irrigate Nebraska's ethanol corn crop alone. And ethanol is basically an agricultural proof-of-work function.
Also even in the totally fake scenario where that water completely evaporates, it's just gonna rain on Iowa a bit more.
You can’t eat your algorithm.
So while pointing to agriculture and saying look there - those guys & gals way more than us it’s a red herring because we need food.
Sure improve efficiency of ag, but a better comp for water usage should be made. Reasonable people tune out.
you can't eat ethanol either, you burn it to create more global warming
That's why they talked about ethanol.
How does it compare to the water used growing alfalfa for export to Saudi Arabia?
"... and keep your hand up if your ethical, water saving stance means you've become vegan."
Odd to hide behind an LLC (LLCs) then.
Real estate is almost always "hidden" behind an LLC or REIT, as recommended as common practice by CPAs and attorneys.
It would be odder if it were directly in Google's business name.
Do I understand the logic here to be "if there's an LLC with an obscure name involved, it must be using too much water"?
To me, the logic is companies don’t put their brand on activities that contribute to tragedies of the commons.
Alphabet could have named the company “google data center 083941, llc.”
But Alphabet deliberately obfuscated its control of the facility.
No, it’s that if there’s nothing to hide, then they shouldn’t jump through so many hoops to avoid basic industrial reporting requirements.
It’s not any one thing, it is the litany of things they do in order to avoid any transparency or accountability. It is undemocratic and contradictory to the idea of good governance of public resources.
If anything, a little sunshine here would probably be the best medicine. Trying to be opaque is what allows conspiracy theories to exist in the first place.
I'm commenting here about the water figure. Are you talking about something else? Because the impact of that amount of water usage is knowable regardless of the corporate structure of the data center.
That’s more of a convenient way to distribute and isolate risks than it is anything nefarious.
You see, that is nefarious. Not taking responsibility is nefarious.
Or to shift blame and hide wrongdoing.
What I now wonder is why the hell they're trying to hide it then. If the numbers are strongly on your side, why hide them?!
Because nobody reasonable believes water use is any issue, but these numbers are proxy for the usual stuff companies don't disclose for usual reasons (ones that rhyme with "competition" and "stock market")?
They sure as hell aren't making it easy for the "reasonable" to take their side when their response to the "unreasonable" is "the numbers you claim are bad are actually good, but you don't get to see them because they go to a different school".
We have perfectly good reasons to hide behind shell companies and ndas! Anyone who disagrees is a hater!
Because hedge funds are going to use those numbers to get better estimates of their business and revenues and use that in their forecasting which means Google loses in the stock market.
Are you regularly filling Olympic sized pools and checking the water usage of an entire town in Nebraska? I would rather see this put into context by comparing it to other businesses normalized by power usage or square footage.
No, I'm not checking the water usage of Olympic swimming pools, because I intuitively understand that they are not meaningful users of water.
I can see how a 0.27% increase in Lincoln’s water usage intuitively seems insignificant, but that was a a low usage data center in a large city. Papillion has 1/12 of Lincoln’s population, yet it’s data center used 42x more water. Assuming water usage was proportional to population it would have increased the city’s water usage ~136%.
It doesn't seem insignificant. It is in fact insignificant. And it supports one of the most popular and widely used consumer products of the last 100 years, so it seems like a pretty incredible bargain.
Tech companies are bundling AI into their base subscriptions and force feeding it to consumers, so I doubt the metrics they are reporting to their shareholders represent real demand or value.
>they are not meaningful users of water.
How so? NOT COUNTING leaks and filling, the average Olympic sized swimming pool uses 1-2 million gallons of water a year.
See EPA Commercial Pooland Spa Equipment Best Management Practices for Commercial and Institutional Facilities https://www.epa.gov/system/files/documents/2023-11/ws-commer...
Yes. GP is literally saying that 1-2 million gallons of water is a pretty small number not meaningful to anything at scale.
It's basically like framing power consumption in watt hours. A million watt hours is tiny at any scale, your whole neighborhood uses millions every day.
Please compare 1 million gallons to the 1 quadrillion gallons in the aquifer, or the 31 trillion that falls from sky each year in Nebraska and get back to me.
Nebraska's main product is corn. Irrigated corn takes roughly 2 acre-feet of water, so the total DC water usage in the article (765M gallon) could irrigate 1k acres of corn. That's a single average-sized farm, not even a big operation. Calling a rounding error might be too generous.
Different state, but: almond trees growing in California consume 1 trillion gallons a year (estimated).
So probably we could grow fewer almonds in California.
How do you get the California water to this datacenter?
In other words, water scarcity is a distribution problem, not a globally finite resource problem.
People shouldn't be latching onto side effect costs of AI data centers like water or energy use, it seems like an abstraction layer above the core complaint. If we don't want AI, or LLMs, or data centers that should be the complaint. Once it is abstracted to water use or similar it opens the door for data to show the actual water use is comparatively small.
Its like arguing about climate change by zeroing in on carbon emissions. The issue people should be raising is wasted energy, over consumption, not knowing "enough", etc. Once the issue is pinned to a single externalized factor the argument lives or dies by that factor.
This article seems to focus on one of the less interesting in terms of water usage data centers (for most of us non-locals) because the paper is from Lincoln and that data center is in Lincoln.
The article's data center used 13 mil gal of water, but as they point out another one used up more than 500 mil gallon.
Seemingly much better article (pictures, better comparisons): https://flatwaterfreepress.org/data-centers-can-guzzle-serio...
Apparently Google self-reported the total 700+ million, so the main thing to learn from redacted numbers is how much smaller all the other data centers are compared to the 500 mil gallon one.
Update: in terms of water usage. Not sure whether 500 mil gallon one is larger or just really water-inefficient. That would be a nice thing for some journalist to research...
Update 2: it seems that Google uses water-evaporation cooling in regions that they don't consider water-stressed. Here's a not-so-great reference (scroll down if the text-fragment link doesn't work): https://www.axios.com/2026/06/03/google-pushes-water-standar...
The concern around data centers looks a lot like a moral panic. They've been around for decades and nobody cared before the backlash against AI. They're like any other industrial building, subject to the same resource constraints and zoning laws. I suspect if the owners of that land built a steel mill, a semiconductor fab, or anything other than a data center with similar power and water requirements, nobody would care.
You are overlooking scale and pace. No one is trying to build more steel mills in the next 3 or 4 years than have been built over the entire history of the steel mill industry so far, with each new steel mill using around an order of magnitude more energy in the same space than steel mills built a few years ago.
I'm not against local governments setting zoning laws and regulations if they're concerned about the pace of growth. Personally, I'm not concerned because I think we can and should invest in infrastructure to keep up with demand. This will benefit all industries, not just AI. My point is that I suspect (I'll admit that I haven't done any research on this) that people wouldn't be against facilities with the exact same footprint if they didn't carry the label "data center" and this is primarily driven by a backlash against AI rather than the resource use itself. In that sense, they're trying to rebrand a political/culture war issue as a zoning or utility issue.
There is sometimes incorrect conflation between what the datacenter is permitted for and what the datacenter actually draws.
> Many articles about current or future AI data centers report the number in the water permit they apply for as the amount of water they actually use day to day. But this almost never happens. [1]
This is from the problematic and often wrong but in this case correct Andy Masley. It's not completely clear to me whether this number is observed usage or a theoretical permitted maximum, in which case the actual usage could be even lower. It could be either.
[1] https://blog.andymasley.com/p/the-ai-water-issue-is-fake?ope...
Why is Masley "problematic" and what's he been wrong about?
He won’t toe the official line. You will always get a vague claim of being “problematic” if you don’t challenge the status quo.
He's done just terrible things, the accusations are true. No one wants to associate with him because of all the bad things he's done. Everyone knows he's a liar, and anyone asking for evidence of it is a perfidious pro-AI ideologue.
He’s a Pro-AI ideologue who engages in the same motivated reasoning he complains about others using. His piece on heat island effects fell over under scrutiny because he just doesn’t have the analytical background to actually do these investigations.
What? I follow him on Twitter, and he's pretty open worried about AI existental risk. He just wants accurate reporting about the situation.
Ok but if you are permitted for some obscene number you could at any point draw that down no?
Why, to flush money down the drain? It would be crazy to go near your limit except in the hottest part of summer.
You can get better cooling performance at the cost of slight water usage or you can use more energy without using water at all.
The greener solution is to use water.
This is entirely dependent on where you are geographically in terms of rainfall and weather. Both water and energy are abundant. The trick is getting them where you need them. So the where is what matters.
Seems like it'd depend a lot on e.g. carbon impact of the marginal watt v.s. environmental impact of the marginal gallon? Might depend on the individual site.
Is there much impact of a marginal gallon? This is evaporative cooling, no? It's gonna come back down as rain and refill wherever datacenter took it from.
Depends on where it's coming from and who else needs water. If the water supply being drawn from was collected from rainwater initially and there's still enough for local residents to also have water, then sure, it's probably sustainable. But if the water is coming from an underground aquifer and there's otherwise not enough rain in the region to go around then it's a finite resource.
I dont like the olympic swimming pool comparisons so I converted to immigrant population comparisons. How many people's worth of electricity and water are the data centers consuming , of course all the analysis was done on gemini since we are talking about google data centers :) here's the output
> Data centers impact local utilities unevenly: while their water consumption is roughly equivalent to adding a small town of 25,000 people, their electrical footprint across the state exceeds the residential power needs of the entire population of Nebraska.
> Together, Google’s three Nebraska data center sites draw enough electricity to power up to 290,000 homes—equivalent to supplying the entire residential power grid for roughly one-third of Nebraska’s total population.
Why do you say “immigrant population”?
Now compare to Nebraska's water usage...
> Agate LLC reports expecting a refund of $55,822,472 from 2025 taxes. Fireball LLC is expecting a refund of $39,171,573.39 from 2025 taxes. Westwood Solutions LLC, the Google-owned data center in Omaha, is expecting a refund of $22,558,881 from 2025 taxes.
Wow!
Regardless of the comments below there seems to be a crazy ideology forming about being completely pro or anti AI. On both sides people are either under or ober reporting how much power these datacenters actually use
kind of confusing numbers here. the 13 million gallon figure refers to one specific datacenter, but you add on five more datacenters and the number grows to... 765 million gallons? The datacenter in Papillion accounts for 548 million of those. Is it bigger? Is it less efficient? not a lot of info apparent in this report.
If those numbers are yearly, they are significant (about what the city of Lincoln uses in a month) but perhaps not troubling. Depends on your perspective.
We're gonna need a comparison chart in number of almond milk lattes and/or golf courses. Otherwise what are we even talking about?
Roughly 40 acre feet of water for the Lincoln DC. An acre of corn uses maybe a couple of acre feet a season. Corn irrigation alone uses tens of millions of acre feet of water per season in the USA.
I don't really understand the hand-wringing about data centers and water. Unlike fossil fuels, when you "use" water, it doesn't go away, it goes back out into the environment. And since DCs use it for cooling, not some industrial process, it doesn't even contaminate the water in any meaningful way.
The one exception is if they're using fossil water pumped from aquifers, which doesn't renew (back into the aquifer that is), but nobody seems to bother to make the distinction.
True if you're using a closed-loop system. But if you're using evaporative cooling (which is the main cause for concern) then the place where the water 'comes back into the environment' may be quite far away (at least in a useful sense).
I think the bigger issue with water usage ( if it's actually an issue ) is when the system is not able to handle the additional demand ( drying up wells, etc. ) Especially during the summer where water supply is already strained in some areas.
Yes water goes eventually back into circulation but that does not help in the place where the system is already at it's limit
20-25% of the water consumed for evaporative cooling becomes contaminated waste water. It is used to clean out the residue left behind by the 75-80% of the consumed water that is evaporated.
The contaminants consist of dissolved solids and minerals that were in the incoming water, plus additives such as anti-scalers, corrosion inhibiters, and antibacterials.
How much additive gets in there? Is it more or less contaminated than normal sewage? A reasonable amount of treatment is part of the water bill.
The excess minerals don't worry me much.
I would love to see a real breakdown of aquifer water usage.
IMO the water thing is a distraction. It gets people quibbling over water instead of asking bigger questions about the data center.
Mostly, but not entirely. It is a huge issue in the Southwest, and it seems like a lot of DCs are being proposed and/or built in areas where water is far from abundant.
Fresh water supply has been on a downward trend for years. We are already looking at a future without enough fresh water, before the new datacenters come into play. The datacenters are being installed in some of the driest areas of the country that were already water stressed. In some cases they can also contaminate local water systems.
I’m not disagreeing with that but it fundamentally feels like a misdirection. It’s easily hand waved away eg “it uses less than a golf course” and yet it’s still a constant media talking point.
Your use of "into the environment" is too broad.
If you pump water out of one watershed into another, perhaps via evaporation and wind transport, then you better expect the downstream users, deprived of their water, to complain.
For example, the upstream users of water from the Colorado use so much that Colorado River Delta is nearly always dry. Or read about the California water wars.
Further, increased competition for surface water raises prices, which encourages additional aquifer extraction.
The problem in California is that competition for water doesn't raise prices. Much of our water supply is allocated based on property rights rather than as a commodity so prices for some users remain artificially lower than the market price. This causes waste and shortages. Eventually we'll have to reform the system but there's no political will to do so.
The water for the datacenter comes from the same place residents, farmers, etc get their water. If there is already not a lot of water, the datacenter can cause running out of water. Sending used water downstream doesn't magically reappear upstream.
In some cases there is no stream involved at all, which is more problematic, but is pumped from groundwater.
> For context, 13 million gallons is enough to fill around 20 Olympic-size swimming pools, and is less than half of what the City of Lincoln reports using on Sept. 29.
Over the course of a year, that doesn't sound terrible.
That's wishful thinking, I know, but would love a breakdown of those metrics into what actually goes into AI inference, training versus just TikTok / YouTube streaming, or just a Google search...To have some perspective. My hypothesis that prolonged video streaming might contribute similarly to AI usage? (But CDN etc might make that tricky to calculate...)
It's not terrible at all. It's dwarfed by common municipal demands, let alone industry or agriculture. It's an amount of water that's basically not worth talking about.
This is also likely insignificant compared to how much water is constantly evaporated by a large lake.
Data centers are ugly and dystopian. We can likely survive as a species without building a single other one. Society and general well-being does worse because of them and what they bring.
I wish people fought gasoline refineries as much as they fight data centers.
From a redaction point of view: I can't believe this is still happening. Surely by now all government and local government offices have a data/records person and policies about information handling. I guess it's not really sensitive info like the name of a government informant. Maybe they intentionally leaked it with a 'whoops, we tried' excuse.
If it really matters, block it out, print it, scan it, then send that.
If it really really really matters, block it out with random extra amounts of width, print it, scan it, edit the page images in paint to move the text to the right of a redaction forward or back a bit so people can't guess the words based on width.
It… doesn’t look that bad. Electricity use is sizable. Water which usually gets all this screaming is, well, look up how much a golf course or a large hotel uses.
It varies pretty greatly, but our golf course uses untreated water from a relatively dirty river that is not used for drinking. It looks like Google is using ground water from the aquifer. These are significantly different beasts.
Concern about water use is the distraction they caused in the media so we'd too busy thinking about fake issues to think about the real issues.
Not a conspiracy theorist, but wait that really sounds effective. Just few months back I was wondering why were big datacenter companies not trying to sway the public that the water usage is low per query or whatever metric.
There will be protests and cases related to water usage of AI, but none of them would hold any ground. And winning the water case would just make the argument in their favour.
Which begs the question of who was the source of the water/electricity FUD?
You don’t have to create unrest, sometimes you just have to let it go unchecked in the right spots.
Flood the zone.
> Just few months back
Just a few months more, and the public was still often making Epstein references.
Excess water use is actively harming people right now.
Seems like a lot of water to me. And what does society get back from google in return?
Yeah, what have the Romans ever done for us?
So what? Water used for cooling doesn't disappear.
Historically large data centers used evaporative cooling. So the water disappeared in sense that it's no longer available in a useful form in it's original locations. That's changing with growing use of closed-loop systems.
A lot of people think it does, surprisingly. People memed too hard about "querying ChatGPT drains lakes" and anyone who contested it was deemed a booster, so it stuck.
I thought some cooling setups aren't closed loop?
https://en.wikipedia.org/wiki/Water_cycle
It doesn't mean the water goes back into the same sources it was pulled from, though
no it doesn't but it doesn't go back to where it was taken from either. So it does affect local aquifers.
This document was created after the Epstein redaction debacle and yet they still leaked redacted data. People will never learn. <facepalm>
If everything was on the up and up they wouldn’t be hiding things behind shell companies and ndas. The water and electric usage is the least of our worries. These things will make cancer alley look like a Sunday picnic. There is a reason executives won’t live anywhere near the hot zone.