
Loading summary
A
The rising inequality and growing political instability that we see today are the direct result of decades of bad economic theory.
B
The last five decades of trickle down economics haven't worked. But what's the alternative?
A
Middle out economics is the answer.
B
Because the middle class is the source of growth, not its consequence.
A
That's right.
C
This is Pitchfork Economics with Nick Hanauer,
B
a podcast about how to build the
C
economy from the middle out. Welcome to the show.
B
Where are you, Nick? You're in Seattle with me, right?
A
I am Seattle, Washington, usa.
B
And it's a lovely spring day atypical for Seattle. Not a cloud in the sky.
A
Yeah, exactly. And today we're going to talk about clouds. And you know, cloud is such a harmless sounding thing, isn't it?
B
It's light, it's airy. Yeah, it's like it's not there. Except actual clouds are real. Like, you know, they're filled with water vapor. And sometimes clouds are good for the world. That's right.
A
What would we do without clouds?
B
Right, but we're not talking about that kind of cloud, are we?
A
We're going to talk about the AI cloud, the data centers that people are building everywhere to power the AI Boom. And that cloud is definitely not ethereal and harmless. The cloud that we're building involves the kind of physical infrastructure that, I mean, think it. I think it's almost safe to say the world has never seen before.
B
Right.
A
Buildings so large that they stagger the imagination, consuming so much power that, you know, it's beyond precedent.
B
I think you're pointing out something which is something we don't really think about in the tech world because we are accustomed to the dematerializing of the economy. Right. That as we in the information age, information is not physical in form, but of course to store it, to transmit it, to use it actually requires physical infrastructure. And the reason why we don't think about is that most of us don't see it. It's the. It's the cables and the fiber that's underground that's stringing along the telephone poles, but also now to feed.
A
Yeah, but it used to also be servers too, right?
B
To be fair, yeah, the servers are there, but we're not seeing them. But we also weren't building them at the massive scale that they're being built to fuel AI. Or as it's now called, not just. It's hyperscale.
A
Right.
B
Just because we have to add that word, hyper, and it doesn't even do it justice in terms of the amount of physical material that goes into it.
A
What's happening around the country is that people are waking up to the fact that their communities are being transformed by the addition of these giant facilities adjacent to them. And that transformation, while it may have some positive effects, also has huge negative effects that the community is asked to bear. Right. For a very nebulous benefit. And today on the pod, we get to talk to somebody who has learned a lot about this. Tim Murphy, who's a national correspondent at Mother Jones, spent what appears to be a big proportion of his life recently traveling around the country, talking to people and witnessing the evolution of this AI boom and how it is physically affecting the places where it is appearing. And it's a super interesting story. It speaks to a lot of the issues we cover in the POD about plutocracy and inequality and, you know, a world in which a few people benefit and most everybody else gets screwed. So with that, let's. Let's bring Tim into the conversation.
C
My name is Tim Murphy. I'm a national correspondent at Mother Jones magazine, and I'm here to talk about my cover story for our May June issue on newsstands now about how American oligarchy went hyperscale. It's in the age of AI we
B
have one of those oligarchs here, though not an AI oligarch.
A
If only.
B
If only. All right, because you're not rich enough.
A
So, Tim, we read with interest your really interesting article where you sort of excavate the physical impacts, if you will, of the AI bubble. Maybe it didn't catch you by surprise, but it sort of caught the rest of the world by surprise. Usually these tech waves happen in bits, not atoms. Right. And this one is happening in atoms as much as bits. And the impacts, I think, are beginning to be felt and recognized. So why don't you just tell us a little bit more about that and what you found in your reporting.
C
Yeah. You know, actually, as I read this and started working on this, I had this classic early Wired story from Neil Stevenson, the novelist, and they devoted the entire issue to it when they were laying a transatlantic fiber optic cable. And his entire deal was he just kind of followed the cable to kind of its end point. And he just wrote about every single thing that went into the physical Internet. And I had that kind of idea in my head when I was thinking about the AI boom that that's sort of being pushed right now. And I wanted to sort of get a sense of, like, what that actually looks like and what that feels like. Not right, you know, not on your smartphone, but in your community, in your Backyard. I think what's been really interesting is that a lot of people's millions and millions of Americans, the way they're experiencing the AI boom at the moment is not necessarily through, you know, through their office, through their smartphone. It is through their community and the reaction to the physical infrastructure of it right in, in the AI hyperscale data centers. And, and so what we've seen starting in around 2023, 2024 and, and rising and rising ever since is, is a sort of a building boom with little precedent in modern times. Sort of evocative of things like the transcontinental railroads of these massive companies competing to see who can build the biggest, most powerful structure in the most places the fastest. And that is having major repercussions in communities across the country.
B
Yeah, it's interesting you started by mentioning fiber, that we had one of these booms, but that was very different because it was invisible, it was all buried and it was mostly being that infrastructure was being put in along existing right of ways. So it was not disruptive to communities. But of course the whole Internet that followed, even though there was a crash in, that was built on that fiber infrastructure, which is still invisible. This is just must be overwhelming in some parts of the country in terms of the amount of physical infrastructure and the disruption in building it. If you could go into a little bit of what it's actually happening at the community level.
C
Yeah, in certain parts of the country, and particularly in Northern Virginia, it really is overwhelming. We had some incredible photos in the piece which you can read@motherjones.com of like a baseball field at a elementary school in Northern Virginia and there's just, just a massive Amazon facility or something like that towering over it. It's kind of everywhere. These communities are running out of spaces where they can actually develop these data centers. And NOVA is kind of a unique place in some respects because there just is going to be a demand and because there is a geographical reason for there to be a massive data center industry because of government, because of fiber optics. But across the country in communities that are being targeted by specul by third party agents for hyperscale sites, you know, people are kind of furious at the, at the demands that are being placed on them and kind of the facelessness that they perceive from the companies that a lot is being asked of these towns and they're not even really getting kind of the dignity of knowing who's asking because everything is being done under NDAs and they're to some extent expecable, you know, reasons for a large corporation to do everything under NDAs, but it hides the corruption. Yes, that, that's certainly a sentiment that you would get at a, at a town hall. If you go to any kind of community meeting in the United States over the last year, people just wondering who the heck is building this thing that is necessitating billions of dollars in, in, in tax breaks. And they are enormous. You know, I mean, they're measured in like units of Manhattan. You know, something like the OpenAI Stargate facility in Abilene, Texas, just outside of Abilene, is about the size of Central Park. And if you drive by it at night, you know, as I did last summer, it feels like you've discovered like a secret base under construction in an alien movie or something like that. They're, they're not some early Microsoft data center in Redmond, Washington. They're, they're the size of airports, basically. And, and so they kind of feel like you've built something enormous and completely different in the middle of a community that's really used to that.
A
Wow. And, you know, one of the things that the world is waking up to, and certainly these communities have woken up to it earlier, is that the impact these things have on electricity prices and water availability and. Because they require so much resources. Right.
C
Yeah. And, you know, there are solutions to, to different extents for, for some of the different problems that are raised by a hyperscale data center. For instance, you know, when they're first kind of under construct of the stories you saw a lot was how much water they use. Well, it turns out you can design them so that they don't use as much water. The way you do that is you use a lot more power. And that's one of the major complaints that you get is just how much power is needed for hyperscale data center. I mean, literally onboarding reopening nuclear power plants like Three Mile island in Pennsylvania Power, you know, a Microsoft data center, the Meta Hyperion Campus, because they're all named for kind of Greek mythology in Richland Parish, Louisiana.
B
But they haven't learned any lessons because they use names like Icarus. I mean, what happened.
C
Yeah, you didn't, didn't get to. That wasn't included in the AI summary of the story.
A
The names, like, come out of a bad James Bond movie. Right?
B
Yeah. Or AI Apocalypse stories.
C
Yeah, I was just gonna say in Richland Parish, Louisiana, they're building two natural gas plants on site direct across, you know, the street from this Central park size data center, you know, right next to a Church right next to this lovely old lady that I, you know, spent time talking to about it. You know, they're adding tons of carbon into the air in communities and they're driving up utility prices. And so now there's a push to build more and more of these power plants behind the meter. But you know, they're opening up all sorts of big questions about who these public resources belong to, what kind of infrastructure we should be prioritizing and actually building.
B
Do you know what the mix is in terms of the power demand between powering the processors and cooling the processors?
C
No, I don't have that breakdown in front of me. But in order to develop these technologies so they're not just sucking up all of the water out of West Texas or in disproportionate number are being built in water scarce areas like West Texas, like Arizona and hot areas.
B
I mean they're, they're building it in the hottest parts of the country mostly.
C
Yeah. And you know, that correlates to you can just kind of build anything in West Texas. Like you can build this nuclear powered Rick Perry startup outside of Amarillo. You could just build things. But because you're doing this in these water scarce areas, you know, you're acting to spend even more money and use even more power in order to maximally, you know, conserve your, your water resources there.
A
Yeah, new industries are sort of the story of civilization. I suppose this is not the first time that big new industries have come into communities and disrupted them. But I feel like this is slightly different because it seems so obvious that this technology is going to benefit a few people and not everybody. It's really obvious how these things will benefit Elon Musk and his shareholders.
B
And assuming they're the ones who win, because they're not all going to win. We'll get to that in a moment.
A
But you know, and for, you know, for your ordinary person, it's cool to be able to use AI to help you write a letter to your lawyer or something like that. But you know, it's definitely not going to put more food on the table or help your kids lead a better life or.
B
But you know, like, let's talk about the way it's sold to communities. These are obviously, if they're, they're getting tax breaks, it's to create jobs. It's always about creating jobs. Obviously there are a lot of jobs created in constructing these massive facilities. What type of jobs are created for these communities once they're built?
C
It's really striking, you know, in, in the take this community in Louisiana where Hyperion is being built, it's one of the largest data centers ever built. I think there's, you know, something like 11 buildings and two power plants and, and whatnot. Requires a lot of engineering and, and, and will take years to kind of come up, come online. So there will be thousands of construction workers over successive waves. And so you'll have different groups of thousands. You'll have the, the initial crew who are kind of the Earth movers and you know, they're hauling in dir from, from all over and, and they've had to set up man camps and everything is an RV park now and, and there's food trucks everywhere to, to feed everybody in this small, poor rural parish. They're going to leave. And then a new group of, you know, several thousand people will come in to kind of put these buildings together. You'll have, you know, even more highly skilled technicians who are actually assembling a cutting edge, you know, Silicon Valley data center. So you'll have many, many people actually putting this project together. And those are construction jobs or, you know, engineering jobs. And those are all inherently short term. When this thing is actually up and running, if it is ever fully up and running to the scale initially promised, you will have a lot fewer people. I don't know the numbers for Meta off the top of my head, but you know, for Abilene, this OpenAI Stargate project, the actual guarantee for number of jobs that they'll have for like kind of a comparably sized hyperscale facil facility is around 100 now. You know, it could very easily be more than that.
A
It could be 200. Yeah, right.
C
It could very easily be more than that, but it's very, very unlikely to be.
A
It's not going to be 20,000.
C
Yeah, it's, it's not going to be the same as like, if you built, you know, an automotive manufacturing facility, which is what was originally what the site in Louisiana was intended for. They hoped that they were going to lure Hyundai by, by turning this cotton plantation into an economic development site. They never got the car company. Ten years later they get a data center. And the difference between those two job profiles is pretty enormous.
A
100 jobs versus 10,000 or something like that.
B
And you mentioned if it ever does get built to the full capacity, it strikes me you use the term cutting edge, these cutting edge data centers, for how long do they remain cutting edge before the technology is obsoleted?
C
Yeah, I mean, you know, one of the tenets underpinning this that you'll See, somebody like Sam Altman talk about is Moore's law, which is not a real scientific law. It's like kind of a article of faith that, you know, high end computing chips will double their capacity like roughly every two years. And so there is this enormous built out, kind of based on like a 2024 or 2025 understanding of what you need to build out. And not just based on your, your sense of like what you currently need, but also just an enormous kind of guess about what the future needs will be or, or even just a deliberately bloated target based on your desire to put together an impressive market sheet or to kind of intimidate your rivals. So there's all this, this kind of sense where like they're throwing out numbers, you know, $500 billion of investment. You know, I think Mark Zuckerberg said to Donald Trump at the White House dinner they had together last year something similar for Sam Altman where they're just throwing out enormous numbers and it's not really clear to what extent they're real. And so if you're in a small town and, you know, I watched a community meeting like this in Wisconsin, and you're being presented with a giant hyperscale project from an invisible third party, you don't know who it is. Well, you do really have to wonder, is this going to be one of the real data centers or is this going to be one of the ones that is maybe just kind of on the books for a few years while we figure out if we need them? And so people really don't want to be the town that's kind of left holding the bag when this all kind of goes away. Or if this all goes away.
A
Yeah, because if the music stops and somebody's built a central park size structure next to your town, what do you do with it?
B
It's very different. Again, going back to the fiber analogy, we had that overbuilding of fiber before the dot com crash, that was part of that crash. And we had all that dark fiber. But you know, 10 years later that fiber wasn't dark anymore. It didn't go away, it was still usable. That's not true of these certainly. At least the GPUs that are in these data centers. If this is a bubble and things stop, what is the value of these facilities to these communities, to anybody?
C
It's a very large paintball course.
B
Well, that's fun.
A
At least it's 20,000 pickleball courts indoors.
C
One of the interesting examples in this town, Menominee, Wisconsin, that comes up at this community meeting, everybody there is thinking about Foxconn, because on the other side of the state, about 10 years ago, Scott Walker dangled billions of dollars incentives for this Chinese manufacturing company to come. And it was going to create 13,000 jobs. And that was hailed as, you know, a godsend for the deindustrializing Midwest. It created maybe 1300 jobs. It was an enormous boondoggle. I mean, it probably brought down Scott Walker and it's sort of now the poster child for over promised tech manufacturing jobs that are going to save your community. And ironically, what ends up popping up in the Foxconn footprint in Racine, Wisconsin is data centers, Microsoft data centers. Because it's not actually the thing that you land when you want to bring in a ton of jobs. It's the kind of thing that you land when you couldn't bring in a ton of jobs, but you still had this footprint.
B
So let's, let's talk about the end game here. Obviously, it's a handful of tech billionaires who are competing with each other to be the dominant player in. We're not talking about chatbots here, in the end, that they're trying to build, they're trying to build actual artificial general intelligence. Right. Obviously, they all can't win. This is going to consolidate down to 1, 2, 3 major players. What is it that the Zuckerbergs and Musks and so forth. What's their vision of the future? Are they our new end of democracy? They're our new overlords?
C
This is something that comes up a lot in their sort of public manifestos of which they keep producing time and again, particularly Altman and Dario Amadai. And it's something that comes up in their correspondence with each other, which we've seen a lot of, for instance, with the ongoing court case between Sam Altman and Elon Musk that originally centers on this mutual idea that they shared that Google was going to create this all powerful AI that was going to control the world. And it was like one guy was going to be in charge of this. And we couldn't afford to do that. We had to stop that. They had to do like a Fellowship of the Ring type type deal. And they had to band together and develop AGI first, basically. And so that's why they created OpenAI. And as you saw over the course of that experience, none of them trust each other. They don't trust anybody else. And so you have this kind of inherent skepticism of everybody else in this industry getting too much power that drives them all to be the one to acquire all of that power because they don't trust their competitors to develop artificial intelligence.
A
Yeah. A bunch of psychopaths competing with one another for power. What could go wrong?
B
Well, I mean, look, if. If one one of your competitors gets too much power with. It's nothing that a couple drone strikes couldn't fix. I mean, I'm joking, but I'm not joking. I mean, I wonder at what point they start declaring war on each other, since they certainly seem to think they're more important than nation states at this point. I don't know it, Tim. It doesn't. It doesn't make me confident about the future.
C
No. And I think, you know, some of that is. Is sort of built on maybe them overhyping their own.
B
Yes.
C
Technology. It's not like we're saying this from Elon Musk's Mars base right now, but it does, you know, it does speak to, I think, the genuine, like, sense that some of these tech founders have, that they really are the person saving civilization and that it really is incumbent upon them to do this. I don't know that that's as true of Mark Zuckerberg as it is of Elon Musk or Sam Altman necessarily. I think Zuckerberg would be happy maybe if we all just had Ray Bans with cameras on them that hooked up to his latest AI model. But I think what you've seen and then over the last couple of election cycles is what's happening in tech with AI is sort of fused with a, you know, political development inside the Republican Party where you have these people who are building kind of a new civilization and tech have joined forces with people who want to kind of save a particular idea of American civilization, and they've found kind of a common ground and a kind of a techno manifest destiny with AI playing the part of the transcontinental railroads. And that political alliance which has, you know, contributed hundreds of billions of dollars plus the Department of Government Efficiency, has had enormously destabilizing impacts on the democratic process on the United States as a whole. And it's something that I think will continue heading into the midterms.
B
I mean, you know, Thiel is open about this. He wants to destabilize democracy. He thinks democracy is a bad idea.
A
Yeah.
B
So is this. I mean, is this. Is that part of the end game?
C
Certainly for. For Peter Thiel, he felt like American democracy was something that needed to be completely kind of broken up. I mean, he had his own alternative ideas. But even with Donald Trump, I think reported you know, it's been reported his fear was that Trump would not break things enough for, for Thiel's own demands. We've seen what Elon Musk getting in and, and kind of treating the federal government like one of his companies. We've seen what that looks like. You know, you take somebody like Sam Altman, who I think he now describes himself as politically homeless, which he saying for kind of political reasons to be a little bit more flattering to the Trump administration. You know, he's somebody who's also talked about sort of how AI will necessitate some deep societal change. We'll have to completely rethink how we tax and how we spend and how we allocate wealth to society. So I think, you know, the people driving this very much do think that they're producing something deeply destabilizing that will lead and necessitate a kind of radical response.
A
Yeah. I wonder if they're counting on swinging from lampposts themselves, which I think is part, going to be part of the end game here.
B
I'm glad you said it, Nick. Yeah, I think, because I feel, whenever I say something, I think, I think
A
it's going to get bad nary and people are going to be mad and they're going to. And they're going to want blood. What's unfolding in communities? What's the trend?
C
The trend is that if you are coming to a community with a proposal for a Data center in 2026, you are going to, you are going to get booed. You are not, you are not going to meet a welcome reception. And now these are still getting built, and most of them are still getting built, or at least most of them are being greenlit. But, you know, something like in the first half of just last year, something like $94 billion worth of data center projects were paused or scrapped based on community opposition. It's just, it's happening all across the country and it is leading to, it is sort of leading up from, from that ultimate grassroots level up into the political level where, you know, Alexandria Ocasio Cortez is now, you know, called for a full moratorium. So is Bernie Sanders. And like maybe you think, well, of course they have. Josh Hawley and Ron DeSantis are kind of vocally anti data center within the Republican flank. And that's sort of going after that sort of J.D. vancy and like Silicon Valley Alliance a little bit. So I think, you know, this has quickly become one of, it's not quite a third. It's not a third rail of American politics. It's something everybody wants to, to touch and hammer home that they don't like these data centers. And Silicon Valley is fighting a real rear guard action against that now. They've started spending a lot of money on advertisements to try and kind of pitch the public in a way that they didn't really pitch them the first time along. Right. You know, because they've been doing this kind of anonymously and, and, and through LLCs and stuff like that. So Meta has now spent millions of dollars on ads sort of after, you know, this vast pitch about how nothing will ever be the same and we're changing the world. They now spotlight a small town in Iowa that has built data centers. And it, and it's got like Friday Night Lights and it's got, you know, old people hanging out at diner and, and the, and the pitch is the opposite. It's build a data center and you won't have to change. And I think that's really striking given
A
that you're now a global expert on all this. If you were going to give a community advice today, what advice would you give them?
C
You know, one of the things that I've been really impressed at and you can watch a lot of community meetings and kind of recoil a little bit sometimes. I mean they can be a real den of conspiracies and certainly you get conspiracies here. But some people are really extraordinarily well read. And you'll watch this and people will be talking about, you know, for instance, they'll be talking about like data analysis on water and power usage that they read about in Bloomberg Business Week or they'll talk about the latest they heard on NPR or something like that. And so I think just actually doing your research on these things at like the credible sources really goes a long way. And that's the type of research and sort of of just showing up that is not often done. You know, if, if you look at, and I was watching old footage of community meetings when this Abilene project was first coming up in like 2021.
A
Yeah.
C
And, and it wasn't with OpenAI yet. It was just like this third party developer saying we'd like to build a data center here and there's a sigh of relief that they weren't going to be building an ammonia plant because I would smell and basically no one showed up. I mean no one had anything to say. Maybe one person like a local gadfly talked and you know, bane the gavel, great, we're going to build it if that had happened in 2025, there would have been hundreds of people there. And so it's that those two things of, of actually being informed and actually showing up is the difference between what's happening in American communities now and, and what might have looked like at the very beginning of this story.
A
Fantastic. Tim, why do you do this work?
C
I really like to go to random towns I've never been to before. But more importantly, I love to dig into a subject I didn't really understand going in and just really get my hands into it and understand the world and be able to relay that information back to people.
B
You know, I think when you look at the long history of economic development in rural and exurban communities, you see this sort of interesting path. It used to be that you'd build some light manufacturing or heavy manufacturing, a car plant, something, and that is something that would generate jobs for, you know, working class people. You can have your farm, you know, and you can have that, you can maintain the farm. People can be working on the farm and then going into the factory to make money on the side. This kind of steady living, doing manual labor, not unskilled labor, it's skilled manufacturing things. Obviously, you know, these weren't always great jobs, but with the, the rise of the labor movement, they became middle class jobs and you could build a very good living and towns grew up and
A
you can build a community around these
B
things, right, because they employed a shit ton of people. And then of course, the deindustrialization of the 1970s, 80s, 90s aughts devastated a lot of these towns that had come to defend on this manufacturing base and to replace that, we got, you know, Walmarts and maybe you were lucky and you got a Costco. It became more a service industry job which didn't create as many jobs and they weren't nearly as well paying, but they did produce a lot of jobs, they did employ a lot of people and they provided services within the community. May not have been great for all the other retailers when a Walmart came in, but certainly for consumers you got something, you got something out of it, right? It benefited, you can make an argument that it benefited the community in some way and you can see why a lot of these communities were desperate to go to Walmart. That would, that was a boon to them, at least as consumers. And now in this third stage of economic development, instead of a factory that's going to employ 3,000 people or a Walmart that's going to employ 600 people and sell you stuff cheap you're going to get this data center. Yeah.
A
Which is as big as Central park, doubles your electricity bills and employs a hundred people.
B
Employs 100 people, most of whom are. They're not farmers.
A
Right.
B
These aren't loads of. They're bringing these people in probably. Maybe it's not like, you know, this is employing local people who are out of work. So you're not getting a job out of this. Once the construction of the data center is done. And you raised that important point about it's actually raising your costs because of the incredible demand on the electric grid on this demand is growing so much faster than we are growing capacity. Especially now with Trump spending billions of dollars to kill wind farms. Yes. But also, I don't know if you know this, Nick, but in a lot of the country, the water utilities have been privatized too. So these aren't public utilities. And so it's raising electricity rates, it's raising water rates, it is raising property values in a lot of these places, just driving out the locals. And then at the end of it, you get this data center that will employ 100 people for in the best case scenario, 10 years, 15 years, something like that. There could be an AI bust that bubble bursts and you're not employing anybody and now you just have this giant facility that's good for nothing. Nothing like where at least before you were growing cotton.
A
Something.
B
Look, you and I have differing opinions about AI. You love it because you think Claude can replace me and nobody.
A
Claude could never be as annoying as you.
B
That's right. Well, you know, they're working on it. Yeah, that could be a feature. You can turn up the annoying.
A
Yeah, the annoying button. Yeah. Right.
B
You love it. I think I'm disappointed. I mean, it's great at doing funny photos and stuff. I think it's less useful right now than people think it is because it can help you with your research, but you can't trust any of its citations. I think it's amazing it can write as well as it does. It still sucks. And it's not the AI that I was promised, which is that real artificial general intelligence where it becomes self aware and then destroys humanity. But we have this differing opinion about how useful it is and I question put that aside. But we can put that aside entirely because that's not what this is about. No, this is about what it's doing to individual communities.
A
Whether I'm finding AI more useful than yours. Beside the point. Right. It's certainly hard to believe that for the ordinary American family, it's gonna do anything useful in terms of their welfare. Right, right. I mean, you know, at least directly. I mean, indirectly, maybe AI will discover a drug that will be beneficial or whatever it is. But is it leading to children being better educated? No. Is it leading to a tighter kn. No. Is it, you know, like, is it. Is it putting more food on the table for the typical family? Absolutely not. Right. Like, and the downsides potentially are absolutely enormous. If there's any. Anything we can be sure of, it's that Elon Musk, Sam Altman, and Mark Zuckerberg will not likely have to deal with the downsides. Right, Right. They will be completely insulated from that again until the pitchforks come and they actually end up.
B
Right.
A
Swinging from the lamppost.
B
You know what, as Nick. And it seems weird because it's high tech and they're building things. It's a form of resource extraction.
A
Yeah.
B
If you are a community with a giant hyperscale data center. Yeah. They are extracting your electricity, they're extracting your water, they're extracting your land, but they're not actually. The profits from this are going to a handful of tech pros. Right, Exactly. They're not really putting anything back, and they're not creating something that you're directly benefiting. The point being that they could put that data center someplace else and you would benefit just as much. Right. From using the ChatGPT or Claude or whatever. So it is a form of resource extraction, and I have a hard time seeing how that benefits local communities in the long run.
A
Yeah, I agree. Well, another day in the life of plutocratic America.
B
Yeah. If you want to read more, we will provide a. A link to Tim Murphy's piece in Mother how the American Oligarchy Went Hyperscale.
D
Pitchfork Economics is produced by Civic Ventures. If you like the show, make sure to follow, rate and review us wherever you get your podcasts. Find us on other platforms like Twitter, Facebook, Instagram, and threads at Pitchfork Economics. Nick's on Twitter and Facebook as well. Ickhanhauer. For more content from us, you can subscribe to our weekly newsletter, the Pitch, over on Substack. And for links to everything we just mentioned, plus transcripts and more, visit our website, Pitchfork economics.com as always from our team at Civic Ventures, thanks for listening. See you next week.
Podcast: Pitchfork Economics with Nick Hanauer
Host: Civic Ventures
Episode: How the AI Oligarchy Went Hyperscale (with Tim Murphy)
Date: May 12, 2026
Guest: Tim Murphy, National Correspondent at Mother Jones
This episode explores the explosive growth of physical infrastructure behind the AI boom—specifically, the rise of “hyperscale” data centers across the U.S.—and the outsized impact these massive projects have on local communities. Host Nick Hanauer and his co-hosts are joined by journalist Tim Murphy, who shares insights from his extensive reporting on the topic. The conversation critiques the social, economic, and political consequences of concentrating power, resources, and benefits in the hands of a few tech oligarchs, while extracting significant costs from everyday Americans and their neighborhoods.
(00:55–05:33)
“The cloud that we're building involves the kind of physical infrastructure that... the world has never seen before. Buildings so large that they stagger the imagination, consuming so much power that… it's beyond precedent.”
— Nick Hanauer (01:25)
(07:01–10:06)
“A lot is being asked of these towns and they're not even really getting kind of the dignity of knowing who's asking because everything is being done under NDAs … it hides the corruption.”
— Tim Murphy (07:49)
(10:06–12:44)
“They're adding tons of carbon into the air in communities and they're driving up utility prices. And so now there's a push to build more … power plants behind the meter. But you know, they're opening up all sorts of big questions about who these public resources belong to…”
— Tim Murphy (11:30)
(14:00–16:39)
“When this thing is actually up and running … you will have a lot fewer people… For Abilene, this OpenAI Stargate project, the actual guarantee for number of jobs… is around 100.”
— Tim Murphy (15:56)
(16:39–19:19)
“If the music stops and somebody's built a Central Park-size structure next to your town, what do you do with it?”
— Nick Hanauer (18:35)
“It's a very large paintball course.”
— Tim Murphy (19:19)
(20:32–25:36)
“You have this kind of inherent skepticism of everybody else in this industry getting too much power that drives them all to be the one to acquire all of that power because they don't trust their competitors…”
— Tim Murphy (22:18)
(26:03–28:03)
“If you are coming to a community with a proposal for a data center in 2026, you are going to get booed…$94 billion worth of data center projects were paused or scrapped based on community opposition.”
— Tim Murphy (26:03)
(28:03–29:47)
“Some people are really extraordinarily well read…just actually doing your research...and showing up … is the difference between what's happening in American communities now and, and what might have looked like at the very beginning of this story.”
— Tim Murphy (28:13)
(30:18–34:25)
“Instead of a factory that's going to employ 3,000 people...you're going to get this data center...which is as big as Central Park, doubles your electricity bills and employs a hundred people.”
— Nick Hanauer (32:42)
(36:36–37:32)
“If you are a community with a giant hyperscale data center...They are extracting your electricity, they're extracting your water, they're extracting your land, but they're not actually...putting anything back...”
— Co-host (36:46)
This episode paints a vivid picture of America’s AI boom as a new, extractive industry that brings monumental disruption—while few reap the rewards, and many bear the lasting costs. The panel emphasizes the importance of civic engagement and skeptical scrutiny by communities offered data center deals, warning that, absent this vigilance, the AI oligarchy’s hyperscale ambitions may leave a lasting, unequal mark on the American landscape.