Data Centers Are Worse Than You Thought

[00:00:00] Speaker 1: data centers. If you live in a small town in America right now, there's a good chance someone is trying to pitch your city council a data center project. Big tech, the likes of Amazon, Facebook, Google, they all need massive compute to keep everything running. And artificial intelligence is driving the need for these data centers. And not all the projects are built equally. Some are hyperscale, these massive, giant computer warehouses that span city blocks. If designed as pitched, they're going to use as much electricity as entire cities. They lower the quality of the water of the people around them. They spike the electricity bills of their neighbors, and they're very, very noisy. They're not pleasant to live next to. And it's become the surprise, I think, political issue of the 2026 election. People fighting back against these data centers. Well, there was one person I know I had to talk to about all of this, and that's Paris Marks. Marks is a tech critic, podcaster. He's got the excellent Tech Won't Save Us podcast. And they've been writing a book about data centers for the past few years that is just about to come out. It's called Hyperscale: The Ambition and Excess of Big Tech's Data Empire. He's been all across the country. He's been inside the data centers. He's going to tell you how hard it is to get in and out of, which is pretty interesting. So you were on top of this before a lot of other people. You know, because there was reporting on this. It took a couple of years to put together. You traveled places. What was it like to watch over the last six months is prepping for this book to come out and data centers become one of the big political issues across the world, and especially in America? It's been wild in a way. [00:01:49] Speaker 2: Obviously, I figured that data centers were going to become a bigger issue, but I didn't realize the degree to which that was actually going to happen, right? Because the book is coming out now in 2026, but I pitched the book two years ago. And even before that, I was working on a series for my podcast called Data Vampires that was looking into data centers even before that, to start to figure out what I wanted to do, what I wanted to look at, what the issues actually were here. And I feel like for a long time, when we thought about the digital services that we use, what we do on our computers and our phones, it just felt very immaterial, right? We logged on, we didn't think about what was powering all of that, other than needing to have our computer plugged in and paying the telecom provider for some internet. And other than that, it was like, it should be fine, right? But now, in 21, 22, I started to hear more of these stories of, say, how a ton of electricity in Ireland was going to data centers, and that kept escalating and the impacts it was having over there. And then to hear about this case in The Dolls in Oregon, where Google actually sued a journalist to try to make sure its water use figures wouldn't come out, right? And so it was like, okay, there's something percolating here, there's something happening. The companies are, one, trying to build out a lot of data centers, and these can have impacts, but also are trying to be quite secretive about what's happening. And then, of course, you know, in 22, ChatGPT comes out. And by mid-late 23 into 24, the ambitions for data center expansion of these companies are just really starting to escalate. And then, as these projects are proposed, as we're seeing more begin construction, through 25, we start to see the public become much more aware of what's happening with data centers, you know, as I'm kind of doing the research for this book, preparing this book, until what the past six, eight months, where it really seems to have just exploded, right? [00:03:40] Speaker 1: Yeah, everyone's mad. [00:03:42] Speaker 2: Yeah. And like, rightfully so, I think, in many cases. [00:03:45] Speaker 1: Yeah, no, absolutely. I mean, I talk to people every week now, it seems, where a lot of it's kind of the same story over and over again. Small town, 20,000 people, somebody that's behind an LLC that was set up six months ago, comes to town, makes a bunch of promises, makes the city council sign NDAs. And oftentimes, the land is something that was going to be used for a park, or had been set aside for something else, or like has people living right next to it. And the people building the data center, well, don't want to ask, answer questions about power, or water, or noise. And then the fight starts, and people get upset. And I've kind of seen this over and over and over again, in the states, it's been really interesting to watch people become politically active about tech in a way that they haven't been for years, and kind of fighting this physical architecture of the future [00:04:36] Speaker 2: that big tech wants to build. Definitely, you know, and if you think about like a lot of the fights that we've had about tech, it feels much less material, much more abstract, right? You know, we're concerned about the effects of social media on people. And certainly, that's like, you know, affecting people in a material way, we're seeing people in different parts of the world get murdered because of stuff that's posted on social media, you know, we're seeing, you know, Facebook kind of implicated in a genocide in Myanmar, there are these things that happen, but it just feels like, you know, there's a there's a degree of like abstraction or something away from, you know, the kind of tangible aspects of it. Or, you know, you think in like the early 2010s, it was a bit more physical then when people were really taking on like the Google glass, the smart glass is something that's coming back today. But even like, you know, laughing at the metaverse, it's like, you know, there's still this kind of distance from it. But now it's like, the tech industry is not something that you're accessing through your phone or through your computer, but it's like plopping down right in your neighborhood close to where you're living. And that potentially is going to have implications for how you live, right for your way of life or your ability to continue enjoying living in your community. And I feel like, you know, that can really easily lead to like a knee jerk reaction of like, oh, this is just NIMBYism and people don't want development. But it's like, no, like, there are rightful reasons when, you know, a massive building is coming and locating next to you a massive industrial facility next to your residential area, and potentially affecting your water quality, you know, potentially creating a lot of noise through the air conditioning units that are on top of the building, potentially having floodlights on all night that are going to be like shining into your windows. Like this is not NIMBYism. This is a real tangible impact on your way of life, not to mention if it is kind of increasing your power rates because of the demand it's creating, or if it's polluting the local air because of the generators that it's running in order to power its servers. These are real issues that can really affect people in a material way. And so for me, it's no surprise that, you know, people have really begun to react to it, where, you know, in some cases, obviously, the data centers aren't even built, right? But people have heard the stories of what has happened in other communities, and they're responding against that, and rightfully so. [00:06:48] Speaker 1: It's interesting because for a long time, it boiled down to some sort of consumer or personal choice. It's like, well, I'll stay off of Facebook, I'm not going to buy into the metaverse, you know, I'll buy this kind of phone and not that kind of phone, I won't get Google Glass. But now the Ray-Bans, I think are a really good example. If you live in certain areas of the country, there's people around you all the time that have them. And you're like, you're going to be a part of this, no matter what choice that you've made. And it's the same with data centers. [00:07:15] Speaker 2: Earlier this year, I went on, I happened to be over in Zambia for a conference. And I was like, I'm in this part of Africa, like I want to go on safari, right? And so I went on safari in Botswana. I was like, you know, I've probably already traveled all this way. I may as well take advantage of it. I had a few extra days here. Go on safari. There's a guy in my little safari Jeep wearing meta Ray-Bans. And I'm like, are you kidding me? You're not even going to get good quality videos or photos of these animals. Like just use your phone or something. I was just so annoyed. Like I wanted to tear him off and like throw him in the lake for the gators or something, you know? [00:07:50] Speaker 1: Maybe he's listening to podcasts. Maybe that's the use case. One thing that comes up in this a lot is the title of the book. Can you define hyperscale for me? What is a hyperscaler? [00:08:03] Speaker 2: Sure. You know, the weird thing is that there's not really like a set industry definition, which is like funny, I guess. Like if you look back to the definitions that people were using, like, you know, 10 years ago of what a hyperscale data center is. And then you look at the ones that are being built today. It's like, I don't think this really applies, right? People used to say, I think it was like over 10,000 square feet and like a certain number of thousands of servers. I think like over 5,000 servers. But like, you look at the data centers being constructed today. And it's like, you know, you can't even base it on that metric, right? I've seen other places say that a data center that's over like 50 megawatts. So it consumes 50 megawatts of electricity would be considered hyperscale. I think that is maybe a bit more of a, you know, modern way of defining it. Because if you look at these data centers today, you know, even on the lower end, you see a lot of the hyperscale data centers, the ones designed for AI and kind of like the modern cloud data centers. You're looking at 100, 150 megawatts as like, you know, a kind of base place for them to be. Some will certainly be lower than that. But then, you know, if you look at the ambitions of some of the companies like Microsoft or Amazon or Google or Meta, you know, they're planning like clusters of data centers that are a gigawatt, two gigawatts. You look at this one Meta's planning in Louisiana, and it's hoping to get it to five gigawatts, or like so it says, right? And this is like, you know, a quantity of electricity that can power like a small to medium on the higher scale, like a large city, right? That's the equivalent of what we're talking about here. The number of people, of homes that would be using that amount of electricity. It's on a citywide scale. And so this is the kind of stuff that they're planning to do. And so then, when you look at the fact that they're building out so many of them across the United States, but around the world as well. And about half of these hyperscale data centers are in the United States. So, you know, because the tech companies are in the US, you know, the US also gets a lot of that infrastructure. But that also means that then there are so many communities that are having to bear the burden of this infrastructure. And for me, that is also why we've seen the opposition explode over the past few years. Like I think that, you know, and we can get into it. Like there are political reasons, there are like, you know, people angry at the tech industry reasons. But it's also like, you know, in the past, the data centers were smaller. So they were easier to ignore. They were in, you know, kind of the industrial part of the city. You weren't paying much attention to it. But also, there just weren't as many communities that were, you know, facing these infrastructures, right? And now that there's just so many, now that they're trying to build them in so many different places, more people are feeling the effects. And, you know, I think that's why there's this kind of broader public movement against it. [00:10:41] Speaker 1: One of the things that's really interesting in the book is that you're looking at other countries as a preview of like, what is going to be built in America. Because like some of these kinds of projects have been realized other places. It's usually like Chile and Mexico. Because I think that there's kind of the speculation America, or at least the tech industry has been very good about spinning. Like, you know, the water uses isn't going to be that bad. We've got the power thing under control. Like, just let us build them. It's going to be fine. Well, we've seen what happens when they get built other places, right? Can you talk about that? [00:11:12] Speaker 2: Totally, totally. And you see it in the States too, right? Like, I feel like the reason that it can really stand out in some of these other parts of the world can just be because, you know, maybe the infrastructure isn't as, you know, advanced, or as robust, or there just isn't enough kind of resources to hand over to one of these data centers. So, you know, you look in Chile, for example, and Google was planning to build this data center in this community called Cerrios, which is, you know, working class part of Santiago, the capital of Chile. And, you know, for context, the Santiago has been basically experiencing drought for like a couple decades now, you know, for the most part, like, it just does not have the water to spare. But the tech industry is going in there and looking to set up a bunch of data centers, including this one Google is going to set up. And when the community looks at, you know, the kind of environmental reports that Google has submitted, it realizes that there is a ton of water that Google is using here, like far more than the local residents would be using. And they become worried that their access to water is going to be affected by that. Right. And so they, you know, basically go to Google and look for reassurances. They go to politicians and look for reassurances and they get nothing. And so they just start opposing the project. They do a local referendum and, you know, a plurality of people vote against the project moving forward far more than that vote in favor of it. And essentially that project gets stalled because people do not want to have it there. And they're really worried that they have running water in their homes right now. But maybe they will have to go back to relying on, you know, water from trucks because the data center is just going to use so much. And we see stories like that repeated in different parts of the world. You know, you can also look at Ireland, obviously a more developed country, but over there it's more than 20% of electricity that goes to data centers, which is like, you know, the United States is around 4% at the moment. So it's not, you know, nearly there. I'm sure there's parts of the country that are much higher as a percentage of their local electricity. Virginia, for example. Exactly. If you look at places like Northern Virginia, but it's like, you know, there they have kind of grid reliability issues, the power prices have gone up. And, you know, as we have seen this escalation of data center construction in the United States, we've started to see more communities facing these issues as well, right? So issues around water access or contamination of water or things like that when data centers are built. Or you look at the number of states now where there are a lot of data centers being built and where people are paying a lot more for their electricity because the demand created by data centers is driving utilities to have to compete for a limited amount of electricity, which is pushing up rates for, for rate payers, essentially, right? So regular people are paying more for their electricity because of the data centers. And, you know, that's not to mention the environmental impacts that can come with the data centers. You know, if we think about local air pollution and, you know, the, the bigger issue, again, more abstract of an issue, but is the climate impact as all of these data centers start building out dedicated gas power plants, you know, for these massive hyperscale ones, because they're like, we can't access the grid electricity because we've used all that up. So what are we going to do now? Oh, let's tap into the gas network and just build a bunch of gas power plants. And let's not worry about the emissions that come from that. It's wild. [00:14:24] Speaker 1: I mean, the power and the water thing are very interesting to me. The power thing, because I don't think people understand that America kind of builds out as it needs it. The grid is just this patchwork nightmare across the entire country. And like, it takes a long time to build the infrastructure and to do it safely. And so the, the current plan is at a large scale, and tell me if I'm wrong, but this is like kind of what I've, I've, I've put together from like writing these stories and talking to people is we're going to kind of sell that we're going to get everything eventually one day down the line on nuclear power and renewables. But who knows how long that's going to take 10 years, at least maybe longer. And in the meantime, we'll take care of our power concerns ourselves. And the way that they're going to do that is by building literally hundreds of gas turbines, and just like running the pipelines, kind of to these data centers, removing them from the grid, and just, you know, bringing back natural gas. [00:15:22] Speaker 2: We love natural gas, right? You know, it's, it's clean. [00:15:27] Speaker 1: That's why it's, it's natural. Nothing that's natural could be bad. [00:15:30] Speaker 2: Well, it's not called fossil gas. It's natural gas. You know, we fall into the industry terms there, right? And I do it myself. It's what I call it in the book. I would say that was the story, probably six to 12 months ago, right? And I feel like the story has even evolved since then, because they were making all of these agreements with like nuclear companies and nuclear power plants. And, you know, the, the assertion was like, we're doing this so we can get all this electricity, and we're going to power our data centers, and it will be clean. And Microsoft is bringing Three Mile Island back. [00:15:57] Speaker 1: Exactly. I think some other people were looking at bringing up one of the plants in, in California back online, etc. [00:16:04] Speaker 2: Exactly. And, and plus building a bunch of like, of these small modular reactors, which are like kind of an unproven technology. No one has built one that works, by the way. But they will do it this time. [00:16:15] Speaker 1: They've been talking about doing it for 10 years, or longer, actually, and it's never quite panned out. [00:16:20] Speaker 2: No, but this time it will work, right? Mm-hmm. Because the, the tech companies say it will. So, so that was like, kind of the, the argument, right? Like, okay, we're going to need the gas now. But look, we had, we're going to have all this nuclear, nuclear coming on stream. I feel like that has shifted where you see a lot of people in the industry now. And like, you know, Eric Schmidt said it to, I think it was the Senator of the House, like one of the committees he was, he was in front of there earlier this year, last year, where he was just kind of like, we need more of everything. Like, yeah, sure. We need, need more nuclear. We need more renewables, but we need more gas too, right? And there was a report recently that something like 189 gigawatts of, of, you know, gas production is kind of in the pipeline in the United States right now for data centers, which is just like staggering to imagine. And so I feel like we've pivoted from, okay, we're going to get to green energy eventually to like, we just need a ton of gas in order to power this because they're also building these like alliances with the fossil fuel industry. And certainly the Republicans are like really down to use a lot of gas and support the, the fossil fuel industry as well. And we've even moved on from like, I feel like the turbines was a discussion like a year ago when we were talking about Memphis and what XAI was doing there and how it was building out this second facility just across the border where, you know, it was also going to be relying a lot on turbines. And there was just a report in the past couple of weeks, I think about a Microsoft one that was doing something, something similar where it also had a bunch of turbines, I think without the proper permits. So now we've gone from the turbine part to just moving on to straight, like, we're just building full gas power plants to power these data centers because we need the stability and reliability of the electricity. And for some reason that needs to come from gas instead of building out a ton of renewables. I guess like land area, like a gas power plant needs less land area than, than renewables. And, and they consider it more stable, I guess. But yeah, that's kind of like where we are right now. And, and it's like, to me, it's, it's wild because if you look like five or six years ago, all these tech companies were trying to position themselves as like, so green and, you know, making all these climate promises and all this kind of stuff. And now it's just like, ah, generative AI is here. We want to compete on that. We don't want to be left behind. Climate's out the window. And of course, our government seemed to have, you know, stopped caring about the climate as well. And so it's just like, all right, you know, full speed ahead on fossil fuels. [00:18:41] Speaker 1: I think it literally is that it's just the fastest, cheapest way to get the power online. I think it's just full stop. That's, that's the only consideration, right? [00:18:49] Speaker 2: Yeah. [00:18:49] Speaker 1: Because this is a race and it's a speculative real estate boom on top of all kinds of other booms and bubbles. And they, they are rushing to just get these warehouses, these GPU warehouses up and get them powered. And like, it doesn't matter what they break on the way. And it does seem to be like with the Microsoft example, they're willing to just do whatever it takes and then they'll pay whatever fine afterwards. [00:19:09] Speaker 2: Absolutely. And, and I also feel like there's, there's a capacity of gas that is looking to be used as well, right? And so they're a great customer for the gas industry. Up here in Canada, Alberta is obviously a major oil and gas producer. You know, they have a lot of gas that they could be exporting to different markets. But literally, they are looking at data centers, and they are making a pitch to data centers to try to get them to locate, located in Alberta, because they were like, if they come here, we don't really need to transport the gas very far. We have like a customer that is very close, that will be, you know, a very large customer for our industry. And that is very appealing to these, you know, to these fossil fuel companies, right? And, and it's the same in many other different parts of the world where this is happening. [00:19:53] Speaker 1: Texas is another great example. Until Abbott put the pause in the governor, there's going to be like 300 coming online pretty quick, which is like an incredible number all in one state. And one of the reasons why is because there's so many shales and so much oil, and they can just run those pipelines like directly to their power plants. There's one in Amarillo that they've already literally done that. If you're looking for the ideal place to build a data center, a hyperscale data center, where would you build one? [00:20:20] Speaker 2: It's a great question. And, you know, I think it would really depend. Like I think that there are different locations that could make sense, right? Like I think you would want to make sure that you're not in a water stressed area, right? So there is the water to provide for the cooling that would be necessary. Because if you look at these data centers that are focused on AI uses, they need the graphics processing units, you know, that obviously has caused Nvidia to like completely take off and go to the stratosphere. But they need they need water-based cooling because they run so hot, right? So first, you would probably want to find somewhere that has the water to provide. It's also very beneficial if you're in a cooler environment, right? So there are times of the year when you can rely on kind of the outdoor air for cooling purposes rather than using electricity or as much water or what have you, right? Which is kind of the pitch that we're seeing being made in Canada and say the Nordic countries. And I'm sure Alaska does this as well, you know, where it's kind of like, we're cold, you know, you can save some money, you can use less resources by coming up here. But then you'd also want to consider, you know, the kind of proximity to population centers, to residential areas, where it's going to be located to make sure that you're considering, you know, the different aspects that come with it, right? Also, like, is that area going to have the electricity to provide without putting a significant strain on the grid? And I feel like one of the things that really stood out to me and that I hope that readers take away from the book is that I'm not trying to say we need to, like, stop all data centers and get rid of all data centers. But rather, I feel like the conversation on data centers has just been defined by the tech industry and what the tech industry wants, and, you know, what needs to exist to serve their ambitions, when actually, I think that we should be having more of like a collective conversation and creating a regulatory structure that asks, you know, what is the right amount of resources or the acceptable amount of resources to dedicate to data centers? And what technologies are those data centers powering that, you know, that justify this kind of build out and this deployment of resources toward this end? Because I feel like one of the things that really stands out to me is we're building all these data centers now on the justification that they'll be needed for generative AI. But then when we look at generative AI itself and kind of what we've been promised that that technology is going to deliver versus what it actually seems to be delivering, it just doesn't make sense to be sacrificing this much for a technology that seems to [00:22:40] Speaker 1: give us the public so little. I think that's like an important nuance to this is the conversation that we're having now, we probably couldn't be doing its current form, if not for a data center, or several data centers, like in the pipeline somewhere. Totally. Like, we all rely on this, but like, you get a say in the way that this is managed, and what happens in your community. And it's interesting, because a lot of, again, a lot of the community people I talk to, like, they do understand this, the person who was arrested for clapping during the data center meeting, they told me, you know, like, we've got this big field that's next to my house, that's where they're going to build it. We have an abandoned industrial park that has the space that is already zoned for this kind of thing. That's just the conversation I want to have. Why isn't it over there? [00:23:32] Speaker 2: You know, it's a different but a similar example, right? Like, I'm from Newfoundland in Canada, and there, you know, we have a bunch of hydroelectricity here, but there was talk about building out more wind energy capacity. And like, the first big wind energy project that we got was to build it in this like, natural, like designated, like protected natural area. And it was like, why? Like, we have so much places that we could put this, why does it have to be there? And naturally, there was a bunch of public opposition as a result of that, right? And it wasn't necessarily opposition to wind energy or renewable energy. It was like, why are you sticking it in this place of all the places where you could put it, right? And I feel like this is the place where like, there can be such a big disconnect from what industry wants and what industry wants to do. And like, what actually makes sense, you know, for the public, right? Where it's not like, we should have no digital infrastructure. It's like, no, I think that there are tangible benefits that have come from the technologies that we have, you know, is generative AI that one? I'm not so sure. But like, you know, the fact that we can communicate so easily and stuff like that, like, I think that is something that we would want to hold on to, right? We don't want to get rid of that. But how much computation do those more positive or beneficial aspects of digital technology require? And then like, you know, what are we kind of willing to sacrifice in order to continue to have those things? And does it make sense to sacrifice so much on the altar of generative AI, when it feels like the benefits have not really arrived for the most part, but the costs are so much greater, you know, in terms of social and environmental costs than, you know, what Sam Altman told us to expect at the [00:25:07] Speaker 1: end of 2022, right? They keep in the States, building them in deserts, which I find just so very befuddling. You know, Meta is building one or trying to build one in El Paso, which is on the border in the southwest part of Texas. There's the one in Amarillo I talked about where they, like you said, they've got a stress water table. Do you think it's easier for them to get the land in the desert? Or do they genuinely think that they can get away with this long term? It just feels like if you've got a warehouse that's generating heat, and you're going to go and build something in a place where, you know, it's going to get up to 120 degrees for four months of the year, [00:25:42] Speaker 2: it's not going to go well for you. No, absolutely not. And I think it's a mix, right? I think it's like, a land question. But I also think that there are other reasons too, right? You know, if they can get great tax breaks down there, then why not? You know, you probably have some like communities that are desperate for investment in jobs that will just kind of accept it anyway, and just hope that they can make it work. I know in the past, I feel like this is less the case today. But in the past, they would sometimes locate in the south in order to have easier access to renewable energy and like solar energy and things like that. I think that has kind of shifted. And that was more of like a conversation for a few years ago, kind of like pre-generative AI moment or early generative AI. But like, you know, you're talking about the United States, like I talked to, you know, some people in Spain, where again, they're, you know, Meta in particular is trying to build this data center in like central Spain, where they have like no water, their taps like run dry parts of the year, like it's very like water stress and that's getting worse and worse because of climate change. And they're like, why are they locating it here? Like of all places, why can they not bring it somewhere else? And again, the local politicians are largely on board, because they're like, we need jobs, we can promise investment, look, this big company is coming in that can do this. And I feel like, again and again, you know, what I have picked up from the reporting I've been reading, the people I've been talking to, is that so often they will go to these communities that, you know, are poorer, do not have, you know, a lot of local industry left or their industries have, have kind of, you know, up and gone somewhere else. And it's easier to, you know, kind of get welcomed into those places because even though you're asking a lot, you're, you know, promising some jobs and some investment and some tax revenue maybe in 10 years. And, you know, people are willing in some cases to, you know, kind of put up with the costs of, of getting that, unfortunately. [00:27:33] Speaker 3: If you've been following our reporting, you know that for a lot of companies, AI has become one of those expenses that can grow incredibly fast without anyone really noticing. One team adds a tool, another starts using a different model, usage goes way up, and suddenly you are spending thousands and thousands and thousands of dollars without even really knowing what you're getting out of it. That's exactly what RAMP is trying to fix. RAMP is the financial operations platform used by more than 70,000 businesses to manage and control spend. Now they're bringing that same visibility and control to AI. With RAMP, you can see spending across AI providers set real-time controls and guardrails by team and model and understand which spending is actually moving the needle for your business. And there's an important distinction here. RAMP isn't incentivized to get you to spend more on AI. They're working with you to help make sure AI spend is going toward the highest ROI areas. So instead of discovering three months later that some tool or model has quietly become a massive line item, you have a way to see what's happening and decide where it makes sense to double down and where it makes sense to cut. If your business is spending on AI, you should be using Token Spend Management by RAMP. Again, we've been reporting on out-of-control token costs and RAMP helps companies get that under control. Stop guessing where your AI spend is going. Get started with RAMP. And if you're new to RAMP, you can get $250 off at ramp.com/404media. Go to ramp.com/404media, ramp.com/404media. I've been trying to pay more attention to the boring everyday stuff I buy over and over and over again, because that's where small change can actually add up. Laundry detergent is a perfect example. You use it constantly and traditionally that means buying another big plastic jug every time you run out. That's one reason I like Blueland. Their laundry detergent comes as pre-measured tablets. You just drop one directly into the drum and wash like normal. No measuring, no mess, and the tablets have a plastic-free, microplastic-free formula. I've been using Blueland and I've honestly been using a lot of the dishwasher tabs, which I really like. They come in this cute little tin. And what I've noticed is they clean really well, no residue. And honestly, my dishes have never been cleaner. Blueland also has a new option called Power Duo laundry tablet designed for tougher loads where you need more stain and odor fighting power. Things like sweaty workout clothes, food stains, pet bedding, or whatever else you've managed to get all over yourself that week. And I like that you're not sacrificing performance just to reduce single-use plastic. Blueland's cleaning products are independently tested to perform alongside leading brands and they're formulated without dyes, parabens, or harsh chemicals. More than 1 million homes already trust Blueland, including mine. Blueland is a certified B Corp and Leaping Bunny cruelty-free certified. Their formulas are EPA Safer Choice certified, and many products have also earned Cradle to Cradle's gold material health certificate. If you're looking to make a small change in your routine, you can get 15% off your first order at blueland.com/404. This podcast is brought to you by NordLayer Browser. NordLayer Browser is an enterprise browser that secures online work without getting in the way of productivity. It gives IT teams full visibility and control over every web-based AI tool, SaaS application, and download across the organization, while making sure employees can still enjoy a seamless browsing experience. Designed for today's distributed workforce, NordLayer Browser helps companies protect access to their internal resources and prevent accidental data exposure. Go to nordlayer.com/browser/404podcast and use coupon code 404podcast10 at checkout to get NordLayer Browser at 10% off today. I've started judging clothes less by what they cost on the day I buy them and more by how often I'm actually going to wear them. A cheap sweater you wear twice isn't really a bargain, but something you reach for constantly because it fits well, feels great, and works with everything, that's where the value is. That's what I like about Quintz. They focus on high-quality wardrobe staples made from premium materials like 100% Mongolian cashmere, organic cotton, and merino wool. They're timeless pieces that are actually designed to stay in your rotation. Their 100% Mongolian cashmere sweater start at just $60, which seems impossible, but you're getting the softness and quality you'd normally expect from a luxury brand without paying the luxury price. And heading into colder weather, Quintz also has premium down jackets and classic wool outerwear. Easy layers you can throw over the clothes you already wear, which for me is usually other Quintz stuff like their t-shirts, their linen over shirts, their pants. I really like as well. Everything at Quintz is priced 50% to 80% less than similar brands because they work directly with ethical factories and cut out the middlemen. A lot of my wardrobe has been replaced with Quintz. I have Quintz sneakers, I have a Quintz rug, I have a Quintz bedspread, I have Quintz pillows, I have a Quintz towel. So the next time that you need to replace something that is out of date or that you're just looking to get a new thing, check out Quintz. Because it's not just clothing, they bring the same approach to hotel quality bedding and bath, kitchen essentials and thoughtfully designed furniture. Find your next fall favorites at Quintz. Download the Quintz app for app exclusive offers or go to Quintz.com, get free shipping on your order and 365 day returns. Now available in Canada and the UK too. And when Quintz asks where you heard about them, let them know it was 404 Media. It's the best way to support the show. That's Quintz.com. [00:33:25] Speaker 1: Beyond building things, so electricians, plumbers, the construction workers, that kind of skilled labor that's going into like the construction of it, what jobs are made when a data center comes to town? Not a lot. [00:33:38] Speaker 2: Yeah. So when you think of a data center, your jobs are in the construction phase. That's the reality, right? You got to build this thing. If you're thinking about these bigger hyperscale data centers, maybe about 1000 construction jobs, maybe more than that, if it's one of the really bigger ones. But then once it's up and running, you're talking dozens of jobs, right? And you're talking maintenance people, janitorial security. You'll have some people who are going around and switching out servers and checking on that kind of stuff. Certainly some administration workers. But it's not like the idea that we have of what a tech industry job is where you're making multiple six figures and you get all these benefits or whatever. These are largely menial jobs that maybe will get paid a bit above average. But it's not this great, well-paid industry where you have like the traditional unionized job that you can live off of or something like that. Unfortunately, they're pretty standard kind of jobs. [00:34:37] Speaker 1: You have been in and out of these things over the past few years as part of this reporting. You mentioned security. What's it like to get in and out of a data center? [00:34:47] Speaker 2: So it's hard to get in. Let's put it that way. Some of the people who I spoke to who have been doing work on this for longer were like, "Listen, before ChatGBT, it was easier to get inside of these data centers." But they really shut the door, especially as the attention on data centers escalated as they started building more of them. So I did contact the big tech companies for the most part, not all of them. And I was largely ignored or rejected. There was a journalist I knew in New Zealand who offered to connect me with Microsoft down there because he had interviewed there in New Zealand not long before that. And literally within five minutes of sending the email to Microsoft's person in New Zealand, it was like, "No, that's not happening." So getting in is tough. Amazon used to do tours. I don't think they still do. But yeah, so getting in is the first step. That's a tough one. So I ended up going to New Zealand to go into a few data centers down there because I knew some people down there who could help me get in and make some connections and whatnot. And so once you do get permission to go inside a data center, then it's like a whole other file game. So I had to send in my passport and identification so they could check me out beforehand. [00:35:59] Speaker 1: You had a background check and a passport screening before they let you in the door? That's pretty wild. [00:36:03] Speaker 2: I don't know if it was a full background check, but they needed my identification so they could know who was coming and all that kind of stuff. And so then I get there. And if you've never been to a data center, they're surrounded by these big gates and fences and stuff like that. They're not the appealing, welcoming infrastructures or anything. So I had to buzz in and he had to check that my name was on the list. And so then the gate opened and I was let in. And then I had to, again, buzz into the front door for him to unlock it and let me into the little lobby area where I had to hand over my passport and they had to check it and make sure I was who I claimed I was. And then the thing that really stood out to me beyond all of that was then when the guy came to bring me through, the CEO of this data center company. Obviously, it wasn't a Meta or Amazon or anything like that. And the one I saw was even quite a bit smaller than the ones that they would be building today. It was still a pretty big data center, right? But every single time that we went through a door in this data center, the guy who showed me around had to scan his card, right? Every door was locked and needed to be unlocked for us to go through. And it would unlock momentarily for us to go. And then other than that, it would be locked again behind us. And yeah, a data center, I think it's kind of what you'd expect, but maybe also a bit more utilitarian than you would expect, potentially, you walk into like the server room, it's pretty warm, because you have all these computers running, you know, you have these like, tall racks that are throughout the, you know, the room, filled with these servers that, you know, hold the computers, the graphics cards, whatever that either store memory, or, you know, kind of enable the various different tasks that they're doing, whether they're running an application, or they're training an AI model, or what have you. But the ones that are that are kind of cooled with air conditioning, you know, you kind of go up to them and like the cold air is running under the floor, right? There's like an elevated floor, and there are these vents like in front of the racks, and the cold air kind of comes out in front of the rack, and then the computers kind of suck the air through. And so you can put your hand on one side of the rack, and it's like, you know, really cold air, and then you put your hand on the other side of the rack, and it's like blowing out like so hot on your hand, and you can see just how quickly it like converts that cold air to heat to, you know, basically take the heat off of the server itself, right? But you know, there are rooms in there that, you know, for example, are holding rainwater if they collect rainwater, so they don't need to use as much of like, you know, the water from the system, there's a generator room, there's rooms where people can like kind of prepare the servers before they actually go into the server room itself to swap things out and whatnot. But yeah, like, it's not like any kind of special or magical place or anything like that. Like it's very industrial, very utilitarian. And the thing that stood out when I visited the bigger ones in like Northern Virginia and stuff like that, you know, obviously didn't go inside. But just the scale, like just the size of some of these data centers now is just wild. [00:39:08] Speaker 1: What was the noise like inside? [00:39:10] Speaker 2: Very, very loud. You know, deafening would be an exaggeration. But like, these rooms are very loud, we had to speak pretty loud in order to communicate with one another. And the room I was in wasn't even full, right? You know, it was partially full of servers and things. [00:39:25] Speaker 1: The ones you saw in Virginia, are they loud when you're outside of them? Or is there like enough noise remediation going on? Or what's the... How did you find them? [00:39:33] Speaker 2: They can be right. And so I didn't notice it as much with some of the ones that I visited, but some you could you could hear it more prominently, right? And it also depends on the temperature outside and stuff at the time that you're visiting, you know, how much it's operating, the makeup of the cooling system actually is. The thing that really stood out to me most in Northern Virginia was like, you know, I was driving around this area called Data Center Alley, right? So in the north, kind of around Ashburn and Dulles and places like that. And it was just like, the number of data centers that I was passing, right? Like, there were just so many of them everywhere. And then the high voltage power cables in order to deliver the electricity to them. But then I would drive through these neighborhoods. And it's like, the data centers were almost inescapable in many ways, right? You know, you would drive through through some of them. And the whole neighborhood would just be surrounded by these walls of really tall trees. Because there was a data center on the other side of the road or the highway or the freeway or whatever. And they wanted to not have to see that, and kind of shield themselves from it. But even then, like the top of the data center would be poking out from above it. Or like I drove through another one that was like, you know, built next to this really nice like body of water, like pond, like whatever you would call it. And, you know, you could you could like walk along this little trail and, you know, sit next to the lake on like a little bench and stuff. But then you look across the other side, and here's this big hunk in Microsoft data center there. And it's like, this is a very appealing view. This is me sounding a bit more NIMBY now, right? I don't want to see the data center. But still, it's like, I don't know, it's wild to be living so close to these things. You know, the final thing I'd say on that is just, you know, when I spoke to like activists in Northern Virginia, one of the things that stood out to me was like, they kind of said, we've had data centers here for a long time, right? And we largely haven't been, you know, fighting them because they were in the industrial areas. You know, they weren't kind of abutting into residential communities. And what started to change around 2018, 2019 is that they started to like escape that containment, right? And they started to get bigger and bigger and started to encroach on residential communities. And you just felt it more and more, right? And the demands were greater and [00:41:51] Speaker 1: greater. And that's when things really changed. Yeah, it's interesting. I've noticed that a part of the pattern that I was talking about earlier, like the first fight in almost every community is about zoning. Yeah. It's kind of creating a new industrial area or rewriting their industrial zoning rules to allow for something like this, because by most standards in most communities, this is an industrial building. And they've got to kind of like skirt around the rules a little bit to get them built where they [00:42:20] Speaker 2: want to get them built for whatever reason. Totally. And you also have, you know, in Northern Virginia, because there are so many data centers there, because they have expanded so rapidly, again, they've had a lot for a long time. You know, obviously, you have the legacy of like the internet itself and its relationship to the Pentagon and the military industrial complex. You have various tech companies that kind of built up there over time that were that were kind of founded and based in Northern Virginia that helped that to expand. And then of course, you know, Amazon, it's kind of largest concentration of data centers, Amazon being the biggest cloud provider out there is in Northern Virginia, like there's just a lot of data centers there, right. But it was it was kind of wild to me to hear some communities in Northern Virginia basically just begging the government to rezone their residential area as industrial so they could just like get the hell away from all this stuff because their quality of life have been affected so much by having so many data centers built around them. [00:43:16] Speaker 1: So like if that happens, they would be able to basically get paid out for their home because it's no longer a residential area and then just leave. [00:43:26] Speaker 2: Exactly. They can sell the land to Amazon or whatever data center company and they can build a data center there and these people can just go. [00:43:33] Speaker 1: So when they're talking about quality of life, is it water and power and noise? Like what's kind of the story they're telling? [00:43:40] Speaker 2: Yeah, so it really depends on the place. Like I think what you hear about a lot in Northern Virginia is the noise in particular, right? Because, you know, having these massive facilities with these air conditioning units can create this just like incessant hum that can be really audible and obviously disturbing to people, right? But, you know, we also see in places like Northern Virginia now that their power rates have gone up as a result of having so many data centers being constructed there and the demand that's created. But I feel like that's less of a reason to seek a rezoning of your community than like, you know, data centers have moved in. Living there is not what it used to be like living there in the past. And unfortunately, even though you really liked your community, you don't want to be there anymore because it has just become not a nice place to live, right? And we see, that's like one example, but we see stories about that. And I talk about some of them in the book that we've seen around the country where like, I can't remember exactly where it was off the top of my head now, I want to say Georgia, but like Meta came in and built a data center next to the homes of, you know, this couple that like, you know, had kind of bought their forever home or planning to retire here, had a nice like forest behind their house. And Meta comes in, tears down the whole forest and, you know, kind of like contaminates the water to such a degree that they've had to replace their appliances. Like multiple times over just a few years, because their water is filled with so much like silt and contamination and stuff. Right. And it's not just their home, of course, it's all the homes around them. But there's so many stories like this that that have come from the construction of these data centers that, you know, you think about the local impacts and then you think about, you know, the billionaires who control so many of these companies and how disconnected they are from those impacts, but also how like, they really don't care, right? Because they are living on this whole other plane of existence and have, you know, they're worried about creating like the AI god in the computer or whatever. Whereas people are like feeling the effects of what they're doing, whether it is the building of their data centers or the products that they're, that they're deploying into the world that are having a lot of negative impacts, right? That they just don't seem to care about. [00:45:41] Speaker 1: Their moral concerns, if they can be said to have any, are for people 100 years from now. Yeah, maybe even a million years from now, right? Yeah, maybe even a million years from now. What's the mechanism by which the water is becoming polluted? Like, how does that happen? [00:45:56] Speaker 2: So it depends on different places, right? And different systems and what have you. So, you know, in the case I was talking about there with the Metta facility, it was, I believe, due to the construction, right? So just the way that they had kind of cleared the land and things like that had caused a lot of silt to get into or, you know, a lot of dirt, whatever it was to get into the water that these people were relying on. And they, you know, basically had their own tanks or whatnot, and they couldn't afford to actually replace it, right? And so they kept having to rely on this kind of dirty water. So if you think about the way that water-based cooling works when it comes to a data center, you know, typically the kind of base model I think that a lot of people think about is like the evaporative cooling, right? And so that is when the data center just takes in a lot of water, cycles it through the data center, you know, to kind of get the heat off of the chips, and then once it warms up to a certain degree, the water evaporates away, right? And there's a little bit that would be kind of non-potable water that would be ejected into the system, but largely that water is going to evaporate. And so that requires a lot of water, because you keep having to take in more to replace what's being evaporated. Now, there's another system that the companies are proposing a lot more recently called the closed loop system, where you take in a certain amount of water, but then you keep that within the facility, and you cool it back down as it cycles through. But that process is much more chemically intensive, right? And so you're relying on a lot more industrial chemicals, coolants, those sorts of things in order to do that. And in some cases, that stuff can get into the water supply. And that can have an issue. You know, Amazon was recently fined basically for having contributed to the water pollution in that area because of its data center. [00:47:42] Speaker 1: How did we come to live like this? That's another interesting thread in the book, I think, that I kind of didn't know, and you did really well charting out is like, why do we live in a world where data centers and the cloud are kind of the infrastructure for all of this technology? [00:48:02] Speaker 2: Yeah, it's a good question, right? And we could talk about that for a long time. But you know, I think that there are a few different ways to think about it, right? Like the first is that, you know, Amazon was really a pioneer here, right, with Amazon Web Services back in the early mid 2000s, when you know, it was kind of building services for itself and realize, hey, we can like sell these to other people, right, and create a new line of business. And that was essentially what they did, you know, first with a bunch of startups kind of building themselves on the cloud, but eventually bigger companies kind of jumping on board as well and realizing that they could kind of outsource their IT and kind of like, you know, in-house computation to a company like Amazon, or later Microsoft and Google as they kind of joined the cloud competition, if you want to put it that way. But there were a lot of benefits that came to them for doing that, right? You can look at, initially, there would have been cost savings that came up doing that as you're getting rid of your own infrastructure and relying on Amazon's. Now, over time, the cloud companies have sold, you know, these companies a bunch of additional services that have caused their bills to balloon. But it also allows them to, you know, kind of shirk responsibility for security, for breaches, all that kind of thing. Like if something happens to their data, it's no longer, you know, this company that maybe is in entertainment or is in some like completely random sector that now didn't have good enough cybersecurity. But it's like Amazon who didn't do the right thing. And now we can point the finger at them. And, you know, the people who hold us responsible now see that we're not the problem. You know, it's a, it's kind of a shirking of responsibility in that way and a continuation of, of outsourcing practices, which kind of has led us to this moment, right? Because you can look back to the 2000s and into the early 2010s as the cloud is growing. But then now, as we fast forward to, you know, the early 2020s, when the cloud has become such a major thing, when, you know, Amazon, Microsoft and Google are such dominant players, the fact that we have, I would argue, and I think this is an accurate argument, the fact that we have ChatGPT and generative AI today is wholly dependent on the fact that the computation that we collectively rely on has been centralized in the way that it has, right? You know, under the roofs of Amazon, Microsoft, Google, and some other companies. And so when it came time to develop these models that form the basis of generative AI products, that was only possible because such a small number of companies controlled so much computation and could thus train the models and collect the amount of data that was necessary to create generative AI, right? To create the chatbots and image generators that we have today. And, you know, you can see that very clearly, right? You know, OpenAI relied on Microsoft Azure, you know, the Microsoft cloud service, other AI companies relied on, you know, different cloud providers in order to do it, but like none of them had the capacity on their own in order to do this on the scale that they're doing it today. And the other kind of, you know, maybe wrinkle that I would add to that is like, the reason that we pursue that then is not because that has to be the way that AI develops, right? Because that's not how AI has to be. Rather, you have a bunch of people at the top of the tech industry who have certain ideas of what the future of computation and the future of humanity should be, you know, one where we merge with computers, where we create the artificial general intelligence, the artificial super intelligence, you know, the computer that can think for itself. And so, you know, if that is your vision of what the future needs to be, what computation needs to be, then naturally, you need to build larger and larger models that require more and more computation to try to, you know, to try to create an equivalent of the human brain and the fact that it can, you know, do so much, right? You know, I would say that's all kind of fantasy land science fiction [00:51:48] Speaker 1: stuff, but this is what they believe. Right. I mean, all of this is going on the assumption that more compute powering, more complicated LLMs is the path to the thing that they're trying to get to, which is not necessarily what's going to happen. Absolutely not. But we all pay the cost in the meantime, right? It's fascinating, you know, in the 1970s, if you wanted to use a computer, you'd go down to the university, and there'd be a giant room full of machines, and you'd say like, this is really wild. And then the story of one of the stories that tech has told is that we've gotten that machine down to just, you know, it's your phone, it's this thin little phone now. But the big rooms full of computers still exist, you're just tapping into them, and they've built them some other place. Yeah, and they're huge now, you know? Yeah, it's much bigger. It's like the entire university campus twice over is now full of the GPUs doing the computations. Exactly. And if you listen [00:52:41] Speaker 2: to what some of them say, like, you know, Ilya Sutzkever, who used to be at OpenAI, people might remember, he was one of the people who was like, kind of involved in the very temporary ouster of Sam Altman back a couple years ago. But like, you know, he has talked about basically having data centers cover the surface of the planet over time, having data centers kind of like, you know, wipe across the face of the planet like a wave, you know? And it's like, this is what these people are imagining the future is? Like, whose future is this for? You know, it's wild stuff. [00:53:11] Speaker 1: Do you think they were kind of, the tech, I mean, was caught flat-footed by the pushback that they've received? [00:53:19] Speaker 2: Oh, I think 100%. You can look at the way that they've responded to the criticism that they got in like the late 2010s, early 2020s, right? And how that really affected them and how they lashed out as a result of it. You know, you think back to like Mark Andreessen's techno-optimist manifesto and, you know, just the types of things that they were saying that the anger that they were displaying at journalism and critics and things like that, because they were praised for so long, and how that deeply affected their politics, you know, their anger at like Twitter and social media because people could criticize them directly. And, you know, obviously the pivot to the right is in some part kind of, you know, we can say pivot, whatever, at least they're kind of open championing of the right and of Donald Trump is some part, I would argue some part kind of tied up in all of that, right? The fact that they were no longer treated as like the wonderful geniuses, but kind of the evil masterminds now. But, you know, I definitely think that they are living in their own bubble and their own world, and they have an idea of what the future should be and how we need to get there. And they think that they are like, what is it like Moses or whatever, who's bringing like the stone tablet down to us mere mortals, right? They all read Nick Land and they've decided they're the prophets of the future. Yeah, absolutely. I feel like there's always been a certain degree of faith in the tech industry, right? You know, it's been not like a religious faith, but a faith in [00:54:38] Speaker 1: technology. Well, you brought up Sudskiver earlier, who there were people during that whole debacle of trying to ask Sam Altman, it came out, I think Bloomberg reported it, that people in the office were saying Sudskiver was like treating the frontier models as if they were gods and sentient and had like, it was like leading chants in the office and had an effigy that they would do things with. There is a certain amount of religious faith born into some of this stuff for these people, I think. [00:55:10] Speaker 2: A hundred percent. The religion is like the technology, right? And I feel like more than ever you see it come out in the AI moment where it really feels like they just have an inherent faith that the technology is going to become, you know, this, this kind of like all thinking, all powerful being that we are going to be lesser too, it is going to control us, you know, we are going to like merge with it or whatever, you know, and we are kind of like building the god and they are reverent toward it, right? And, and, you know, I think for some of them, there's a certain degree of deception in, in kind of seeding these stories. But then as you talk about Sustgiver, I think there are other ones who like really believe this kind of stuff, right? Like really have this intense faith in the technology. It's not like a, you know, a traditional religion as, as we understand it, but it is certainly like, you know, it feels like a new religion or something like that. I remember back in the day when early internet, there was like the church of Google and stuff like that. And people would be like, see the search engine, you can ask it everything. And like, it was a par, it was a parody kind of thing, like the flying spaghetti monster and those sorts of, you know, joking online kind of things. But now it feels like they've just really kind of bought into creating their own religion and their own gods. And, and this is the direction we're going. And somehow one of our biggest allies in the fight against AI is the Pope, which is just like, well, it's a weird timeline, but it's wild how like, you know, obviously the, I'm not going to say the Pope is perfect. Like, you know, don't, don't take this as being like a ringing endorsement of Catholicism or, or, you know, the papacy or something from me. But it is wild how like, I don't know, he's so like openly opposed to AI. And it's like, listen, if we need to form a temporary alliance with the church on this, I'm kind of down. [00:56:57] Speaker 1: There's a couple of years ago now, maybe like two, he was involved in one of the first big, like, it was one of those AI viral moments where the picture was good enough that it fooled a lot of people, it got me at first blush, the puffy jacket, right? [00:57:11] Speaker 2: Right. Yeah, yeah. Well, that was the previous Pope, right? [00:57:13] Speaker 1: Yeah, it was the previous Pope. Yeah. So it's like the Catholics, maybe they do have a little bit of an understanding about what's at stake with, with truth and lies and generative AI. [00:57:21] Speaker 2: They've had to, you know, get ready quickly, but they've also been following this stuff for, I was talking to, I was talking to a monk, who, a French, Franciscan monk, who, and I'm blanking on his name off the top of my head, but who was like an advisor to Pope Francis on, on AI stuff. And he was telling me that when he told like, you know, his, his, uh, group of monks or whatever, I don't remember the right terms, unfortunately, um, that he wanted to go and like, you know, do studies on technology and ethics of technology and stuff like that. Like this is, you know, what, probably two decades, 15, 20 years ago. Now they were like, they didn't really get it. They were like, why is that what you would focus on? And he was like, now it's like they're all turning to me to like help them with these questions because it's like so relevant, you know, um, it's fascinating. [00:58:11] Speaker 1: So what do you make of their counter arguments? The, now that big tech is facing this backlash, uh, they're caught flat footed. We've got Mr. Wonderful going on TV and saying it's the Chinese. [00:58:23] Speaker 2: I feel like I need to constantly apologize that Canada has, has kind of given him to you guys, [00:58:29] Speaker 1: and now you need to deal with them too. We're deeply upset about it. I still feel like they're kind of, nothing's really sticking for them on this. Like nothing is working. Why? And do you think there's anything they could say? [00:58:40] Speaker 2: I know, I don't. Right. I think it's a mix of a few things, right? I think first of all, the arguments are ridiculous for the most part. First, you have like the, the Chinese like propaganda one. And like, you know, the Chinese are just kind of coming in and trying to turn everyone against us and all this kind of stuff, which is just like foolishness, right? Like everybody knows that, everyone can see through it. It makes no sense at all. And I also feel like people aren't even like as negative on China as they were a few years ago, right? Like I, I feel like the actions of the United States the past couple of years have made China look not, not as bad as it once did. So I don't think that one works. Then you also have like the data center, the data industry lobbyists, um, kind of saying like, Hey, you know, that problem where we're apparently using a lot of water, not actually a problem. And I also don't think people are buying that one, right? Because they're, they're trying to like, look at like the 10,000 feet level, and basically say like, Hey, nationally, data centers aren't using as much as like golf courses or agriculture. So see, they're not a big deal. But it's like when you're talking about water use, as we've been talking about in this conversation, we're talking about local issues, right? And so it's like on a whole, and remember, data centers are still a top 10 water user, like as an industry in the United States. So it's not like it's insignificant. But when you look at the places where they're locating, there are a lot of those that are water stress regions, right? That don't necessarily have the water to provide to a major data center like this. And so that is a problem. So in that sense, I don't think the arguments are working because I just don't think they're good arguments. But then the other piece is that I think there are a lot of people who are opposed to data centers for a lot of reasons, and some of it will be the tangible local impacts. But I feel like another of the things that's driving this is one, a broader opposition to AI and degenerative AI and the effects that we're seeing from that, that again, they also don't want to address. But also an opposition that has grown in the past few years to the tech industry itself. Because one, I feel like we don't get the benefits from their products that we were once sold or that were once available to us because they've eroded over time as they've sought to suck more profit out of them. But then the other piece is that, you know, certainly for people more in the center and on the left of the political spectrum, seeing the tech industry cozy up to Donald Trump, and certainly even some like never Trumpers on the right, right, would very much turn them against these billionaires. And I feel like even just hearing more about the types of things that they're saying the past few years, you know, their opposition to democracy and to people's rights and all this kind of stuff, it's like, no wonder people are turning against them, right? And so then you try to convince them that like, oh, the only reason people don't like data centers is because of like Chinese interference. And you know, we actually don't understand how water works. It's like, what are you guys talking about? You know? [01:01:30] Speaker 1: People hear the data centers next to them, you know, they hear it. [01:01:34] Speaker 2: Yeah, go build one next to like Sam Altman's mansion, or like, go plop one on like Mark Zuckerberg's island in Hawaii or something. I'm sure they'll react really positively to that idea. I remember this story, Mark Andreessen, right? You know, his big thing, it's time to build, we need to be building, we need to build more housing. And then I can't remember the outlet, I think it was maybe Rolling Stone, published a story about the fact that, you know, he lives in this kind of like wealthy enclave, Atherton in California, and was fighting tooth and nail to stop housing development anywhere near his like set of mansions or whatever that he owned there is, right? And it's like, build, baby, build, like, let's build everywhere, just nowhere near me. [01:02:20] Speaker 1: Just let me have my little enclave and we'll do whatever we want to the little people. Exactly. That's their mindset, even though they wouldn't admit it, you know, openly. Paris, thank you so much for coming on to 404 Media and talking to us about this. The book is Hyperscale. It should be out the 20th of October, 20th of October, 20th of October. It is excellent, gets my recommend. If you want a great view of how like how all of this came to be and perhaps where it's going. Check it out. Thank you so much. [01:02:50] Speaker 2: Absolutely. It's been a pleasure to talk to you. Thank you. [01:02:52] Speaker 1: 404 Media is journalist founded and listener and reader supported. If you'd like to support us, please go to 404media.co and sign up. You get unlimited access to all of our articles in an ad free version of this podcast. You also get access to a subscribers only section of the show where we talk about a bonus story every week. The podcast is produced by Alyssa Midcalf. Another way to support us is to just leave a five star review of the show. It really does help. Or, you know, you can just tell someone about us. We'd love that. This has been 404 Media. We will see you again real soon. 504 Media. We'll see you again real soon.