Limitless: An AI Podcast - The AI Water Use Controversy: How It's Completely Misunderstood

Episode Date: January 21, 2026

It's time to debunk myths about AI's environmental impact, focusing on the exaggerated claim that data centers drain water supplies. We reveal that Colossus 2 in Memphis uses just 346 million gallons yearly, comparable to two and a half In-N-Out locations. Today we confront flawed statistics and highlight commitments from tech giants like Google and Microsoft to improve sustainability.------🌌 LIMITLESS HQ: LISTEN & FOLLOW HERE ⬇️https://limitless.bankless.com/https://x.com/LimitlessFT------TIMESTAMPS0:00 The Water Consumption Myth3:41 Comparing Data Centers and Burgers6:22 Understanding Water Usage8:01 The Cooling Process Explained11:03 The Source of the 1.7 Trillion Claim13:19 Misleading Comparisons15:21 Local Impact of Data Centers16:58 Future Solutions for Water Use19:16 Key Takeaways from the Discussion22:01 Wrap-Up and Next Steps------RESOURCESJosh: https://x.com/JoshKaleEjaaz: https://x.com/cryptopunk7213------Not financial or tax advice. See our investment disclosures here:https://www.bankless.com/disclosures⁠

You've all heard the headlines. AI is draining our water supply. Data centers are stealing drinking water from the communities. ChatGBTGBT is literally drying up the planet. It sounds terrifying and it's also completely fake. Today we're going to bust one of the biggest myths on the internet and walk through the actual numbers which reveal something crazy that the world's largest data center uses about as much water

as two burger joints. We're going to walk you through exactly how much water water is needed for the biggest status center in the world, Colossus 2, and why the majority of water used is only a fraction of the water used by your local golf club that your friends go to every recant. So part of the feedback that we've seen as we become more of a presence, a platform in the world of AI, is that there's a lot of narratives that try to take down progress. And the newest and hottest topic has been the topic of water consumption to the point where when I talk about with my friends about what I'm working on and what I'm interested in, that frequently comes up

as the first rebuttal. It's like, this is horrible for the environment. It's using so much water, so much energy. And this episode, we're going to focus on the water, particularly around the public perception versus comparing it to the reality and how far off it really is. We're seeing on the screen a few headlines of people from Utah, very viral publications and headlines that have been talking about this. But the reality is, is it's simply not true. So we're going to methodically, and in a very fun way, kind of dissect how wrong this actually is, starting with the book that kind of spawned at all. The author's name is Karen Howe. No pun intended. She really, like Karen, Karen seems like a very fitting name as an author called the Empire of AI, claiming a Google Data

Center would use 1,000 times the water of an 88,000 person city. Studies projecting AI will consume 1.7 trillion gallons of fresh water by 2027. So these claims are a bit outrageous, but this has been the narrative that people have been using as they discuss their rebuttals against why AI should exist. And this simply is not true. And I guess this is kind of where we got skeptical. We were like, okay, 1.7 trillion gallons, a thousand times more than an 88,000 person city. These numbers are outrageously large, but they're being peddled around as if it is truth in these large publications. The thing is, like, we work in the AI space, Josh. So trillion dollars, trillions, we're kind of like used to it. At this point. Trillion here, trillion there.

Yeah, right. But what I wasn't used to is a trillion gallons of water. And I was, The first thought I had to my head was what is this compared to? Like, I can't quite comprehend how much water this is and how much water is used in other industries that aren't AI adjacent. So I started getting skeptical, and I came across what I think now is the most important bit of journalism done on data centers by these guys at semi-analysis, they are a crazy team of researchers that kind of like dig into. to all the boring, nitty-gritty, hard data center stuff

to bring us the facts about lots of different things, including how much water is consumed at the data center level. And the revealings are super interesting, but what I love most about this is Josh is that not only is it really informative, it's also hilarious. Because they've compared... I love this post so much. Right.

They've compared data center water usage to both. restaurants, specifically in and out. So for some context here, they took the largest data center in the world, which is Elon Musk's Colossus 2. It's a data center based in Memphis, and it is the first data center to reach a gigawatt of compute. We actually mentioned this on yesterday's episode. If you haven't seen it, definitely go check that out. It's super cool. So we're talking about $18 billion worth of GPUs here, Josh. So as you can imagine, it's probably using a lot of water, right? So the math that they uncovered was the most revealing part. So Colossus 2 has an annual water footprint of 346 million gallons of water per year. Guess how much the average in and out store consumes per year?

A lot. 147 million gallons per year. So that... Per store. Per store, sorry, yeah, per store, yes. So that means that the largest data center in the entire world, currently today that everyone's complaining about consumes the same amount of water as two and a half in and out stores. Why aren't we protesting in and out? This is super fascinating to me. And it gets even better than this. If you think that's funny, it gets even better. So basically, the report looked at tokens per burger. So we were able to get a metric for how you can actually justify output. Wait, wait. Run that unit by me again. We are breaking the news right here on Limulus. We are measuring

water efficiency by tokens per burger, and a single burger's water footprint is about 245 gallons of water. That equals 2.7 billion AI output tokens, which roughly equates to one burger being equivalent to using Grok 30 times a day for 668 years. So the numbers are just absolutely astronomical. Before I'm done, wait, one more thing, one final conclusion to this. There are over 400 in and outs. There's so many of them. And there's what? Maybe 10 AI data centers. So the scale and the magnitude at which they are wrong is just so astronomical. I found it really funny. Like once you actually get into the numbers, you realize this is so not a problem that it's not even funny. Wait, dude, so this is, you're telling me, so this is incredibly

clickbaity, basically. The figures are astronomically wrong. And if you compare it to just a casual restaurant, we're talking about burgers only here, by the way. But don't know. We're not talking about the fries, sides, and everything else. All the milkshakes. God knows how much water the milkshakes use. We're just talking about the burgers, including their restaurant output and the supply chain for this. So the fact of the matter is,

this problem isn't really a problem. It's actually massively overblown. But what I'm curious about, Josh, is like, okay, is this a realistic take? Like, I know that data centers use a lot of water. How are they using it? And, like, how much of this is renewable versus is actually burned and evaporated into the air?

and we'd never get it back. Like, what's the pros in there? Yes. Yes, yes, yes. Okay, so the truth, they are using lots of water. There's hundreds of millions of gallons per year that are going through these data centers. Some of it gets lost. A lot of it gets preserved.

And the way they do it is there's two types of cooling. There's dry cooling. And then there's adiabatic cooling, which is the process where air cools down without exchanging heat with its surrounding. So if you remember the old iPhone episode that we did that talked about vapor chamber, you can imagine scaling a vapor chamber to the size of an industrial data center. And that's roughly how the adiabatic cooling. works. So it evaporates the water. The water leaves the system, and that's where they lose about

267 million gallons per year. And then the second loss function is the flush and discharge. So one thing that I learned in preparing for this episode is that, I mean, there's a lot of mineral buildup in the water that they use. And 67 million of those get discharged as waste water per year. But those are the two ways they do it. It's dry and it's adiabatic. And there are promises in 2027 and 2028 from a lot of the major AI labs to decrease this waste to about 95%. Currently, it sits about 90%. So 90% gets recycled, 10% gets lost. Those numbers are going to increase incrementally

until about 2030 when the number is actually net positive. Okay, so it sounds like the classical way of thinking about how data centers get cooled is you run a bunch of water in pipes through all these different GPUs, and it removes the heat from these GPUs so they are able to perform at optimal levels, right? And part of this water evaporates, never to be seen again, and some of this water gets kind of recycled over and over again. And the adabatic system, I think is what you said, is kind of a hybrid of both of these things,

but most importantly, it's more of a closed-loop system. So we've kind of got like the majority of the water being renewed. So there's a comparison between water being renewed and water being consumed, which means lost forever. Do I have that right? Yeah, so you could think of a closed-loose system, going back to the vapor chamber, if a vapor chamber is 100% efficient at being closed loop, where water reaches the processor, it heats up, it evaporates, as it evaporates, it dissipates

the heat. This is that at scale, although with some sort of a loss function at the end, where some of that evaporated water does currently exit the system. It sits now at about 10%, and that's where that loss comes from. But I'm guessing it's not 1.7 trillion gallons of water it's a far cry from the 1.7 trillion number. And then there's a second aspect to this that gets a lot of criticism, which is the actual power generation, how much water gets used in the generation of energy through these turbines that are natural gas, some solar. And the answer to that is, in the case of classes, too, 0%.

There is no meaningful water consumption of power generation at all. The entirety of it comes from cooling the GPU system, which is closed loop and done with water. So maybe we just go into where that 1.7 trillion number even came from. This is the source of a lot of the narrative that we've seen play out over the last few weeks. Exactly. So I'm showing all of you folks who are watching a scientific paper from UC Riverside titled Making AI Less Thirsty. When you say scientific, you have to do air quotes. Yeah, sorry, sorry. Sudo-Science. From this university called Making AI Less Thirsty, uncovering and addressing the secret water footprint of AI models. Now, this paper is the source of a lot of the clickbait headlines and TikToks that you watch on.

line. It's this one number. 1.7 trillion gallons of water will be consumed by AI data centers alone by 2027. That's around the corner. We're talking about next year here, right? And so a bunch of people kind of threw up their hands and started protesting data centers because they were like, that is so much water. We humans aren't going to be left with enough water to consume ourselves. But here's what this study actually says. It claims 1.7 trillion gallons of water are used for withdrawal. That's a fancy term of saying, recycled. So imagine like the water being taken, used to cool down the systems, the GPs that we just mentioned, and then recycled again and again and again.

So it's not net new water we're talking about here. Josh, do you want to know the actual consumption that is being like permanently removed, the ones that we should be trying to protest? Yeah, well, I mean, based on that, it's what? It's at least a full order of magnitude off than what is projected. No. It's 100 to 158 billion, not trillion, billion gallons of water. That's 10%.

Actually, it's less than 10% if you take the lower bound of the reported number. So the water that's actually being used by the data centers is only kind of like 10 to 15% of that. And okay, so some people then go, like, what about the drinking water? Like what percentage of that is being affected from the study that was being made? It's only 3% of the headline number. So everyone took 1.7 trillion gallons of water and assumed, all of that water was being wasted, never to be seen again, used again, and it was pulling from other resources. When that strictly isn't being true. So the point I'm making it, I'm going to reiterate it again,

the water, the 1.7 trillion gallons of water isn't being consumed. Think of it like this. Imagine diverting a river to run through a mill. We do that today, right? And it flows back into the river. That's called withdrawal. That's the 1.7 trillion number that I'm pulling out here. So most of the power plant water is being reused over and over again. So that's only the first major kind of bit of pseudoscience that we needed to bust. But there's this second thing. And Josh, you mentioned it earlier from ironically Karen Howe, who wrote this famous book called The Empire of Air.

Let me get this up here so everyone can see this book. It has been rated over 1,300 times on Amazon. but more importantly, it has been quoted directly by The Economist, the New York Times, and so much more. For this specific stack, it says Google's data center will use 1,000 times the water of an 88,000-person city. Guess what? It was off by a factor of not a thousand times, 4,500 times. Josh, can you run me down these stats? Because it's just insane, dude. No, that stat actually makes me sick. It's hard. So instead of comparing it to Google, we're going to start with golf courses because that's, it's just, if you want to come out of it, you got to start with the golf course. The average golf course, 312,000 gallons of water. Desert golf courses, 1 to 2 million. A Google Data Center in Virginia is 400,000. Now, what does that 400,000 give you? That powers Gmail, Google Drive, YouTube, the entire G Suite for billions of people around the world with the equivalent of 1.2 golf courses worth of 1.5.com. water for billions of people. Google's global data center that powers four billion accounts

equals 43 golf courses. And in the state of Arizona, there are 370. So if you're, if you're comparing apples to apples here, that is the Lufx comparison. And it's just so off. This essay that was reported, this book that was published, it's off by like several orders of magnitude. It's not even close. I mean, I'm looking at some of the things that they got wrong because someone did a breakdown of this book that's been quoted so many times and that's behind all these headlines. The book reported 5 million liters as the city's annual water use, but that was misconstrued. What she actually meant was 5 billion cubic meters for understanding here. Leaders and cubic meters are very different things. Very, very different things. So she confused

liters with cubic meters, which is already a thousand X error. But then she said the data center would actually use 3% of the municipal water system, not 1,000x. That's where we led to the 4,500x factor off that she was. So basically, it is an absolute, like, incorrect piece of reporting that has been spread by some of the most important and popular media publications that we've seen. Josh, I want to kind of take your golf course thing a little further because I actually quite like that. Let's just... Okay, we'll keep going on the golf courses. Okay, so what's the number one state that's been getting a lot of protests of our data centers. Arizona. Arizona, big time.

Let's just look at the golf courses in Arizona. Arizona has 370 golf courses. Each golf course consumes about 1 to 2 million gallons of water per day. So that's around 400 to 800 million gallons per day for Arizona golf course, right, for all the Arizona golf course. If you compare that to data centers that consume 905 million gallons or rather no point 1-2% of county water. That's like just in Arizona alone.

So golf courses in Arizona, 29 billion gallons per year, and a data center, the biggest one in Arizona, or actually collectively, all of these, actually, my correction,

is just 0.1.2% of that, 900 million gallons per year. So it's obvious that there's just a lot of misinformation out there, and I think it's really important to just bust this wide open completely. It's just wrong.

And part of me thinks, that a lot of the hatred, if I'm being honest, Josh, I think comes from people kind of equating AI to enriching people that they don't maybe like. Like, I get it. Like, AI will be used as a tool to enrich billionaires even further. It's so wasteful. I don't use it. Well, people don't realize is the average activity that you do on a weekend or the average bit of food that you might consume, buying a burger, 668 years straight of using grok 30 times a day. People don't do that that often. It's just, it's important to level set, in my opinion. Yeah, there are a lot of valid criticisms. This is unfortunately not one of them. But there are some instances in which it does,

or it has in the past, actually affected their localized effects where very small towns have actually felt an impact of this. They date most recently back to 2022. There actually hasn't been many sources recently that have determined that it is making a problem. But there have been some instances. The first one was in Oregon where a Google data center consumed 20, of the town's local water supply. There's another one from 2019 in Virginia. Virginia famously, that's where a lot of the internet data center runs. I think a majority of the data center is run from these Virginia AWS servers.

It consumed 63% of Loading County in Virginia in 2019. But since then, there really hasn't been that much of an interference with the local water supply. And there have been solutions proposed from the companies who are most responsible for creating that strain. and that is Google, Amazon, Meta, and Microsoft who all committed to be water positive by 2030 using new cooling technique.

So that vapor chamber that we were talking about earlier will be done at an industrial scale and will actually be able to preserve 100% of the closed loop water supply. And I think that's going to be a really big deal. That paired with direct-to-chip cooling is also going to make a big difference. If you remember our Vera Rubin episode

where we talked about how the new shifts are cooled, the actual cooling temperature, if I remember right, It was like 115 degrees Fahrenheit that it could be. So now you could actually cool these chips with hot water. It requires much less. Air cooling is getting a little more interesting. There's a lot of solutions coming along the way

that will make this a lot more resource or a lot less resource intensive. Cool. So if I were to summarize some takeaways for this myth-busting episode, there's a few that come to mind. Number one, the numbers just don't support the panic that people are putting it out there. The fact is, AI data centers currently today, and it might change later, use a fraction of what golf courses, agriculture, the t-shirt that you're wearing consumes to produce. So even if we tripled AI water usage today, it would still pretty much be a rounding error. Number two, the context matters. You can't confuse 1,000 liters with 5 billion or whatever the number was cubic meters of water.

that is super important and comparing water consumption for AI data centers and your average burger joint might just be the comparison that you need to kind of like set you straight and be like okay well maybe this isn't that important

going forward. And then the third thing which I think is kind of underspoken about quite a lot is I think if I were to guess we're moving towards a world where we end up actually using less water for data centers

right? Part of it is due to like the different systems like the adiabatic system that you mentioned Josh but also like I think a lot of these AI companies are going to start building water recycling plants to kind of push that 90% water renewability figure that we mentioned earlier,

much, much higher. I think Colossus 2 and Elon is doing that right now for Colossus 3 actually. They're building out a water recycling plant. So I think overall this is a nothing burger, pardon a pun, and we're going to look back on this. We're going to look back on this in the future and realize that we're consuming water in much faster ways in so many other industries that we aren't currently protested. So keep quiet, eat your burger, and let the AI flow.

So if you had to guess what the next narrative would be, that has a negative spin on it. Do you have any ideas? I think my answer is going to be energy. I think they're going to start to converge on the correct argument, which is the energy consumption on a localized scale, starting to actually impact the cost per kilowatt of the average person's home. And how does that get offloaded? Well, a lot more natural gas turbines, a lot more solar panels.

And the process of scaling that up is happening, but it's happening slower than the scale at which they're consuming. So when you take a gigawatt data center like Colossus II, that is consuming the equivalent collective output of San Francisco localized to a small town in Tennessee, there are impacts there that are real. It's just a matter of time until those kind of get uncovered

and then get dealt with. I mean, they are dealing with it quickly, but there is a real strain happening on some grids that are localized to where these data centers exist. I agree with you. And I actually think the numbers that will be quoted on headlines about that specifically will actually be closer to home and to the point because it's simple enough to kind of scale new water techniques to kind of cool stuff down. We reported on a previous episode, I think like two weeks ago, that they're not even using cold water anymore. They're using warm water, 45 degrees Celsius or Fahrenheit, which is super one. I think it's like 90 degrees Fahrenheit to cool these systems down. I think it's a different game with energy where we actually do have limited constraint. It takes so much more work and expertise to scale that.

And we're going to have to tap into like town supply or city supply. So I agree with you. I'm looking forward to kind of like unpicking that one in the future. Yeah, we'd be good at this. We should design the next eye up against our own industry. I think that would be much better to consult with the experts prior to doing this next time. That's hilarious. But yeah, I guess that concludes the Mythbusters episode on the first

one that we'll be dealing with, which is the water consumption and the new metric that is burgers or tokens per burger. And in the case of our tokens per burger metric, the cost is very low. And I don't think this is anything to actually worry about. That's the end of the episode, but I'm actually curious whether you guys enjoy this. We hope you guys learn something new from this. Josh and I kind of went back and forth on this, whether we should do this episode. We realized like a myth-busting series could be really cool because there's just a lot of myths and false claims out there and we face it every day. We try and unpack it, spend all of our time figuring this stuff out. And I love that show growing up. It was great. Yes, same. Actually, dude,

maybe we need to come back with some glasses for the next myth-busting episode, maybe a trench coat, like a fedora, maybe we really lean into it. A very magnifying glass. Exactly. But yeah, if you found this informative, if you found this interesting, and you aren't subscribed to us, which apparently is around 70 to 80% of you, please do so. It's a percentage that is far too high. It's way too high. It is actually almost as high as these false headlines that we keep seeing about water usage for data centers.

So if we're describing you currently, it takes two seconds. Please subscribe. Please turn on notifications. If you're listening to this on Spotify, Apple Music, or wherever the hell you listen to this on, please also do the same and give us a rating. It helps us out massively and puts our videos out to way more people so that we get more eyeballs on this and we can keep producing better videos for you.

I think that's it, Josh. Anything from you? Yeah. And just a small reminder about the newsletter. Today we just dropped a new piece that coincided with the roundup, which was a weekly roundup of the five most important, noteworthy things that you want to be informed on. We post that twice a week once every Wednesday, once every Friday. One's a thought piece, once a recap.

So you can join 100,000 other people who are also subscribed getting the info prior to these episodes dropping. And I think that concludes it. wraps it up. So thank you so much for watching as always and we will see you guys in the next episode. See you guys.