Plain English with Derek Thompson - Why Americans Hate Data Centers

Episode Date: August 11, 2026

It seems a universal fact that AI has become an indispensable part of the American economy. At the same time, opposing AI has also become one of the most influential causes in American politics. Nowhere is that tension more obvious than in the fight over data centers. Across the country, communities are organizing to stop the massive facilities that power artificial intelligence from being built in their backyards. In today’s episode, Derek talks with writer Jasmine Sun, who traveled through the Midwest to investigate the growing animosity towards AI infrastructure. Why do Americans hate data centers? Plus, if this trend continues, what happens when one of the most important forces in the U.S. economy runs headfirst into local politics? Subscribe to our YouTube channel here:https://www.youtube.com/@PlainEnglishwithDerekThompson If you have questions, observations, or ideas for future episodes, email us at [e-mail]. Host: Derek Thompson Guest: Jasmine Sun Producer: Devon Baroldi Additional Production Support: Ben Glicksman Learn more about your ad choices. Visit podcastchoices.com/adchoices

One of the things I try to do with my coverage of artificial intelligence is write and record sentences that everybody can agree with. This is surprisingly difficult to do. Sometimes it seems almost impossible to do because if there's one thing I found about AI, it's that it's hard to get a large group of people to agree on the simplest of facts. Take something like the following claim. AI is terrible at writing. You might agree. You probably agree. Much of the time, I agree.

For many people I follow, nothing is more obvious than the flat, lifeless, robot at a math conference prose of AI. But according to a previous guest, Georgetown University neuroscientist Adam Green, many experts in blind tests prefer AI writing to human writing. Even MFA judges in some blind studies with degrees in creative writing prefer the prose written by computers to that written by people. But despite the controversy that surrounds basically everything you can say about AI these days, I want to list three humble facts. Fact number one, artificial intelligence has never been more important to the economy. Tech's share of total U.S. investment recently rose to an all-time high.

On August 3rd, the Wall Street Journal reported that, quote, the AI boom is transforming the American economy beyond recognition, end quote. The economist Michael Pierce told the journal that AI is responsible now for one-third of the country's GDP growth. If AI blew up, the entire economy would feel it. Fact number two, while AI has never been more important for the economy opposing artificial intelligence, has never been more popular. Negative attitudes toward AI are higher today than any previous period of polling, according to a variety of surveys. And fact number three is that those. two facts are related. If AI unpopularity keeps rising, it will have a major effect on not just

AI, but the entire economy. Hundreds of groups around the country are coordinating now to block the construction of data centers. New York and Texas have announced a data center moratorium. So as I see it, AI is becoming utterly indispensable to GDP growth and market returns at the very moment that it becomes the single most efficient applause line in American politics. Those trends are on a collision course. The most striking thing about opposition to AI and data centers is that the folks building artificial intelligence don't seem to understand that opposition at all. Most of these people live in and around the Bay Area in California, where practically no data

centers are being built, by the way. They don't have an on-the-ground view of the politics of AI and data centers in the places where they're trying to break ground. But today's guest does. Jasmine Sun is a writer who recently went on a reporting trip to the Midwest to find out what's really behind the AI revolt. What are people actually saying? What are their motivations? Why is opposition to AI becoming the biggest applause line in politics?

And what happens if the anti-AI forces succeed in stopping data center construction entirely? I'm Derek Thompson. This is plain English. Jasmine, son, welcome to the show. Thanks for having me, Derek. Great to be here. So you recently went out to Michigan and Wisconsin to report on the opposition to data centers. I want you to tell me a little bit about what motivated this reporting.

trips, you know, people are trying to build data centers all over the place. Why go to Michigan and Wisconsin specifically? Yeah. So I had been seeing throughout the spring this sense of a data center backlash across the U.S. bubble up. I had been seeing first stories about the compute buildout, how big it is, how much money these companies were investing into building data centers, and then simultaneously maps and news articles showing how people were protesting data centers, how they were introducing local moratoriums, Bernie and AOC at the national level talking about data center moratoriums. And I was just really shocked, I think, and surprised at some of the intensity of the opposition and the polling, for example, like in Michigan and Wisconsin,

which are purple states, states where both Republicans and Democrats are very competitive and also rarely agree on anything, you'll see something like a 70-30 opposition to data centers, which is much more bipartisan and much more sort of one-sided than solar farms or nuclear or almost any other issue related to development that you can find. And so I was really curious about what is driving the data center backlash and what is sort of building this and motivating this new political coalition that's forming against AI data centers. And I felt that the only way I could actually learn what was going on was to go on the ground to a bunch of places where data centers were being proposed, where they were being fought, where they were being built. So I end up visiting

four sites and talking to people, workers, officials, activists in Michigan and Wisconsin. It was also a good time to be there because they both have or had these like very interesting primaries going, Democratic primaries going on. The Michigan Senate race, which Abdul L. Say, I had just won the Wisconsin governor's race where a DSA candidate, Francesca Hong, is pulling ahead. And AI and data centers have become an important part of both races and the state. of them. So I was also sort of hoping to understand whether these two states and these local stories might end up being a bit of a bellwether for the politics of AI at a national level. You live in San Francisco. You know the people who work on artificial intelligence. You write

about the people who work on artificial intelligence. What do people in the Bay Area think that Data Center opposition is all about? And to what extent did you're on the ground reporting validate or invalidate some of those assumptions? In San Francisco, I would say that the default assumption or caricature of somebody who hates a data center is that they're kind of a TikTok-brained idiot.

They've watched too many short-form videos and they've eaten up too many myths about these things sucking up the oceans, about maybe they're emitting like radioactive waves or something like that. I think the fact that the world water meme and the sense that, oh, AI data centers just use way too much water. And the primary

risk in harmed communities is how much water they use, which I don't think the research proves out is the primary risk that data centers pose. The fact that that, however, has become so memetics and so prevalent on social media has, I think, given a lot of Silicon Valley folks the impression that data center opponents are just highly misinformed and they are just worried about the wrong things. And maybe there is some sense. that people's opposition to data centers is also about their fear of AI, whether that's AI job loss or other ways of not wanting AI in their communities, which I think has something to it. But honestly, what I found when I was on the ground was that the anti-Data Center opposition

was so much more about the way that residents lost trust in their local officials, their local governments and the way that big corporations were coming in and interacting with them. I actually almost never heard people talk about, say, Open AI, the customer that the data center was serving. I almost never heard a resident or an activist talk about specific AI harms like job loss or deep fakes or whatever. What I much more often heard was, how could my mayor sign this NDA and keep this project secret from us?

How could my city council sell out to this gigantic corporation and bring. in this huge development with unclear impacts without any public input. I think fury at both the lack of transparency, the sense that backroom deals were being made and that these big sums of money were involved in that and sort of corrupting people's small-de democracy at home was much more the sense that I got when I was on the ground. I want to recircle something that you said because I think it's so important. When I'm on social media, on Twitter, following the arguments among people who are building AI, I'll often see them make this fact check. Well, they'll say say there's this popular claim about data centers using all of this water, but in fact,

data centers use much less water than is commonly assumed. But that fact check is often assumed to do the job of proving that anti-data center politics is based on a myth. And what you're reporting, I think, does so well is it shows that the actual on-the-ground opposition to data centers is so much more complicated than just a single factoid about water use. And we're going to spend most of this interview, I think, unpacking the many, many layers of local data center opposition. Before we get to those layers, you go to the Midwest, you see these data centers up close. Tell me what's inside a data center for those who don't know

and what's outside a data center for those who haven't seen it. Like, what is its function and what does it look like? Data centers are the most visible physical manifestations of the AI boom. They've existed for a while. Like during the original internet boom, you started to see smaller data centers used for literally processing or housing data. But data centers today, the primary thing that they do is they train new AI models or they serve inference, i.e. when you send a query or a question to chat to chaty or another chat bot, that query has to go somewhere. The model has to process it, do a million complex mathematical calculations, and then send it back. So inside these gigantic buildings, which often span hundreds of acres, live all of these chips and servers that process

these queries and train these models. And they have all of these interconnects between them that require incredible amounts of compute and electricity to make sure that the speeds are fast, that all of these chips are talking to each other. There's big, gigantic fans inside. So the AC that requires the water and the cooling systems comes from the fact that the chips run super hot and tons of AC is required to keep these things cool. And so when you drive past a data center, it often takes even like a minute and a half going at 70 miles per hour at free way speed. These things are gigantic. They're kind of like these gigantic flat warehouses. They're windowless. They're often gray or beige. They're very ugly, frankly. They're often surrounded in a web of

transmission lines that have been built to hook up the data center to the grid. And I do think that there is a role that the aesthetic unpleasantness of a data center plays and the reason that people don't like it. You write that data center opposition is not just about any one thing. It is, you know, I thought of reading your piece as like, I imagined a delta into which many rivers are flowing. And that's what data center opposition is like. And so I want to talk about at least four of those thematic rivers. The first is anti-AI sentiment. I mean, before I read your article, I would have thought that anti-Data Center politics is all about the fact that Americans writ large do not like AI.

how much of the opposition to data centers that you heard about was just exclusively rooted in negative feelings toward artificial intelligence and the AI companies? Almost none of it. I think that some people mentioned AI at some point in the conversation. It was sort of like the cherry on top of this is, and we're building all this for a technology that I don't even like that much, that I'm not even sure is good for the world, good for my kids,

good for my community. But for the most part, the activists who were showing up at city council meetings, the other folks who were residents in the community who are like, why is this gigantic building here? Why is it being built? They weren't focused on the risks that AI posed. In fact, we almost just never heard people even say the words, chat GPT or Open AI or Anthropic in these hours-long conversations. To be clear, I don't want to make it seem like people were excited about AI. I think the fact that they they didn't like AI very much, made the tradeoff not seem worth it. People were more open to new development when they are at least seeing some of the societal benefits that maybe a solar farm or a car plant might bring. And AI seems less useful than say either of those things to a lot of people. But with the exception of some more maybe like ideologically motivated, a state or national level activist who had gotten involved, I would say that most people were much more concerned

with material impacts, with government process and transparency, or with a sort of general distrust of corporations coming into their communities rather than the specific technicals of AI and what was being built inside these things. So if narrow antipathy toward AI played a surprisingly small role in anti-data-center politics, something that played a surprisingly large role is just broad distrust of companies and of government. Let's start with distrust of corporations. Tell me a little bit about Mount Pleasant's experience with Foxcon and how this serves a little bit like a microcosm to help us understand how these small towns relationship with corporations in the last few years shapes their relationship with data centers going forward. Yeah, so honestly, this was one of the most fascinating things to me was that so much of the antipathy towards data centers had been primed by very bad experiences that the,

folks had getting screwed over by corporations, often working hand in hand with their local officials over the last couple decades. And so it's even hard for me to separate out, like people don't like corporations and people don't like government. The story I heard was much closer to an oligarchy or a dark money in politics or a sort of people are getting bribed and sold out type story, where the idea was that my local officials are colluding with or having the wool pulled over their eyes by big corporations to screw us over. So to go to Mount Pleasant, this is one of the sites that I visited. Microsoft has built to hyperscale data centers, the fair water sites in this town of Mount Pleasant.

I think that has 26,000, 28,000 people in it. So it's not super big. It's a little bit outside of Milwaukee. And Mount Pleasant is interesting because the site that the data center now exists on used to be land that was industrially zoned for a big Foxcomplant. In 2017, right after Trump's first election, he was making these big plans about bringing back high-tech manufacturing to the American heartland. Yeah, all of those jobs, all of his factories went to China, but I, Donald Trump, are going to bring it back.

And one of his flagship plans was to bring in Foxcon, the Taiwanese tech manufacturing giant, into places like Mount Pleasant, Wisconsin, to build factories. Foxconn started a huge bidding process, kind of a lot of folks might remember Amazon HQ2, where they started shopping around to a bunch of local communities, including Mount Pleasant, hey, like, we're thinking of building this factory here. We're going to bring in 13,000 jobs, good paying jobs, making flat screen TVs. What kind of tax breaks are you going to give us? What kind of subsidies are you going to give us?

Also, make sure you sign all these NDAs so you don't know what anyone else is giving. And so all these towns, which are trying to figure out how to rebuild industry after the decline of American manufacturing, start getting very anxious about like, oh, shoot, like we need these jobs really bad, we need this tax investment really bad. How, like, let's promise the world to these companies. And Mount Pleasant did promise the world. They ended up taking out over $100 million in loans to build out infrastructure. They set up a TIF, which is a sort of tax subsidy district that is for Foxcon to exist in.

They eminent domained a bunch of land that was very controversial. And after all of this that the small town did in order to invite Foxcon in, they finally win the bid. And not only are local residents furious at the fact that this was ND8 and their city council was keeping mum for so many months and being very shady about what was going to happen, they knew something big was going to happen, but not what. But then it turns out that the contract was in the deal was never as firm as it had originally seemed. Over the next couple of years, Foxcon keeps pulling back their plans saying 13,000 jobs, actually, maybe just let's start with 1,000 jobs. Flat screen TVs, let's start with like some smaller screens. And basically,

the size of the investment that Foxcon ends up making is much, much, much smaller, a tiny, tiny fraction of what was originally promised. And so all this land had been cleared, all these loans had been taken out to build out roads and grids and whatever for a economic project that literally never showed up. And a large lesson that people learned was not only Foxcon screwed us over by telling us they were going to do this and not, but how come my local officials? How come my mayor, my city council was so naive and so uneducated as to offer this company, all of this from our community, without a reassurance that they were actually going to show up? And so much of the frustration in Mount Pleasant and in Wisconsin, we heard this story from so many

people, was not just about the corporations, but also about the asymmetry and sort of the lack of negotiating power that small towns had against corporations, which means that when a data and our developer shows up and they too are like, hey, we're going to make a $2 billion investment in your small town. Why don't you set up a tax district for us? Please send this NDA. Let's not talk about it. It sounds just like the Voscon story all over again. I'm sure many listeners hearing along, just like I reading along, sort of bumped at the idea of NDAs, non-disclosure agreements between corporations and local governments. How does that work? Why are NDAs being used at all? And how is that spooking people?

It's really strange. I mean, so these non-disclosure agreements are, were very common. I would say that data center developers would have local governments signed during the negotiations phase. And, you know, I think part of this is that AI companies are used to doing a lot of stuff under NDA with their own employees with any partners they work with. They're pretty paranoid. And it's kind of a normal thing in Silicon Valley. But further, I think that there's also a reason where NDAs allow, allow you to, for example, play communities against each other because no one knows the terms of the deal. And like one thing that I heard from multiple local officials I talked to was, you know, if we didn't offer them this tax break, maybe they were going to go to Sheboygan up the road. Maybe they were going to go to Caledonia over there. And so it creates this sort of, again, information asymmetry where because no one's able to talk, no one knows what the terms of the deal are. And this allows data center developers to extract more from local communities. It also sort of keeps land prices stable, right? Because like as soon as people know that data center developers have a lot of money,

As soon as you know that it's a data center that's coming to town, when they start buying out, say, farmland in the area, the prices are going to go way up. And so the NDA is also a way for them to protect their side of the deal there. But again, NDAs in government are really unusual. Like, solar farms don't show up with an NDA in hand. And one thing that I heard repeatedly from even the most pro-built, pro-data center people who I talked to was they all regret the NDAs. Every single person I talked to thought the NDAs were terrible. idea, no government ever should have signed them because as soon as you do that, the whole thing looks much shadier, much sketchier than it is. One story that I heard from local officials like

Caledonia trustee, Holly McManus, was signing the NDA and the existence of these NDAs meant that local governments couldn't get ahead of social media. This was something I heard of a lot, was that because you can't actually really keep a secret in these places, as soon as a contractor knows, as soon as some worker knows, as soon as one city counselor knows, rumors are going to spread. People start to get the sense that, hey, who's that showing up at my farm and asking me to buy the property for millions of dollars? And if your government has signed an NDA, people start asking questions and you can't answer them. And so not only do you look very shady as a local official, but also rumors are going to start spreading and it's going to get distorted. And so the NDAs

have been kind of a big, I think, mistake that a lot of folks regret. And you even see now that, say, Microsoft has promised to stop doing NDAs writ large with local governments. And I've heard from from other data center developers that they are also considering pulling back on NDAs because it's really just made the backlash situation much worse. You've described very beautifully the way that the relationship between citizens and government and big corporations plays into this narrative of power asymmetries. There are powers that be that are trying to shove some deal down our throats and we don't want it. Partially or in some cases because we just had this experience with a GM plant

or a Foxcon plant a few years ago and that didn't work. out very well for the town. There's one other element here that I want to make sure that we touch on, which is that if local citizens are absolutely sure that artificial intelligence is the next big thing, is the next automobile, and they might say, okay, yeah, put a data center here. This thing could live here for the next century. But if there's any doubt that AI might be a bubble, then you're going to have more fears that this data center construction project is going to be a stranded asset that not only doesn't create jobs, but also blights the landscape for many years. Did you talk to people who were worried about this stranded asset risk?

Oh, absolutely. And this is also, I would say, a surprise for me coming from more of a San Francisco in Silicon Valley atmosphere is in San Francisco, no one thinks AI is a bubble. People are like, there's not enough compute. If there's any problem, it's that like we have way too much demand for AI relative to how much compute capacity we have to serve those requests. And so the idea of AI being a bubble is not even really considered here. Whereas, you know, I go out and I talk to people, whether, again, like, whether it's activists, whether it's even workers working on the data centers, people don't really understand the technology. They are not really seeing these tremendous transformative benefits in their own lives.

For a lot of folks, chatGBT is like an email writing widget or like it helps you make memes. It's better Google. And so it's not exactly that it's useless, but that their experience of the technology does not really fit these trillion-dollar valuations, right? And so people kind of have this sense that something about the story economically is maybe not adding up. Maybe they're not sure it's a bubble, but they don't know if it's a bubble. They're seeing news headlines that it's a bubble. They're feeling like this thing is kind of useful, but it's not going to replace the entire economy, as these people say. If that's the case, if there's even a 20% risk that the AI bubble pops, what's going to happen to?

to us? Are we going to get left holding the bag? Was sort of one of the phrases that I heard over and over. Like in Wisconsin with these tax incentive districts, sometimes, yeah, these data centers are paying, you know, tens of millions of property taxes a year. That's great. But oftentimes they don't hit the city general fund until 10 years in when sort of all of the local improvements around the data center have already been completed. If you think that you're not even going to get the property taxes until 10 years, 20 years

in, and what if the project ends up being smaller than expected? What if the AI company that is serving that data center goes out of business because this is a really fierce race and who even knows who's going to win the race, then is that going to become a stranded asset? So many people wondered about this. And again, we're sort of in the deindustrialized Midwest where people are intimately familiar with what it looks like for a factory to die and leave town. When I was in Janesville, Wisconsin, the exact site of the data center is this brownfield where there's a 100-year-old GM factory that was sort of a cornerstone of the community for the entire time. when they went bankrupt during the financial crisis, they never finished the cleanup. There's $30 million of hazardous waste and contamination and forever chemicals in the land. It means there's just this 250-acre lot in the middle of town that is so contaminated that they have not been able to sell the land to any new commercial developer because nobody wants to clean it up. And so when you have these blighted areas, when people see what happens when a company

shows up and then leaves without cleaning it up, and they see this high-risk, high-tech industry coming in and making these promises again. It's going to take four years to build the data center, 10 years before the taxes hit. People really start to wonder, are we even going to see any of that? So if the first theme is anti-AI sediment, which you said probably plays a smaller role than most people think,

and the second theme is the fear of distrust of big business and government working together in a way that's going to screw over local populations in the long term. The third factor that I want to introduce here, the third theme, is media. You know, data center opposition occurs on the ground. People are gathering in the physical world at a political rally or an anti-data center rally.

But a lot of it happens on local Facebook groups. Yes. And those local Facebook groups are talking to each other and they're on Facebook, which means that in the middle of chatting with someone from your local anti-Data center Facebook group, you're also seeing memes that tend to be national memes. And when we think about, like, what are the national political memes of the moment? They are anti- oligarchy memes. They are memes about asymmetries of power, politicians enriching themselves, billionaires,

becoming trillionaires while the working class struggles with affordability. Can you talk a little bit about the way you think media, even social media, might play a role in fomenting or thickening data center opposition. Yeah, I've been working through, and I'm still kind of working through what role I think media plays, right? Because on one hand, because of the conversation we had at the beginning about these AI water use things, like sort of causing Silicon Valley to not take the opposition seriously, I worry about thinking, oh, yeah, everybody's just reading fake news, and that's why they oppose this thing. Like, it seems clear to me that it is not all fake news concerns.

Like, there are real reasons to be worried about things like electricity rates going up in your area. At the same time, it is so clear, as you said, that, like, the internet is a huge part of this. So in Wisconsin, the comedian Charlie Barron's who has, like, I think, like, 1.6 million Instagram followers. He's, like, a very successful Midwestern comedian. He started getting really into data center activism. He started hosting events with organizers in communities across Wisconsin, speaking to the local news, becoming one of the most vocal advocates. And so one thing that say the mayor of Port Washington was really frustrated by was he was like, you know, we would hold these council meetings and they're flooded by outsiders, people who don't live in

port, who have seen a viral Instagram video from Charlie Barron's talking about this. And that's why they're here, but this isn't even their community. Another thing that was really interesting was hearing from the Village of Mount Pleasant, where they mentioned that in part because of the Foxcon fiasco, they did not send an NDA with Microsoft for the data center. And they held, they they said upwards of 60 public meetings discussing this project until from when it was started in 2023 to date in 26. And the vast majority of these meetings went empty. Everyone was fine with the project.

Again, all of the land and the tax subsidies had already been cleared. It was only when, according to them, the data center backlash went national when it became kind of a big political issue in the last six to 12 months. That was when locals started seeing that stuff on Facebook, on social. social media and saying, wait, should I actually be worried about this thing? I wasn't worried about it before, but now that I'm connecting the data center in my community to whether it's these big issues about AI, about the oligarchy, or whether it's just reading stuff about the environmental impacts, that was when people really started to get upset. And so what I saw in these communities

was both in certain ways the data center fights were hyperlocal, and yet they were being amplified by national narratives that were sort of giving folks a way to connect their lives. And so, we're sort of giving folks a way to connect their local grievances, like this thing is noisy and construction is annoying, to a bigger narrative about, and in addition, maybe AI's a bubble, maybe AI is going to take all of our jobs, maybe these companies are bad, why are they doing it in my backyard? And so you have this, like, extremely interesting interplay going on. The fourth category I want to talk about is environmental concerns. Like, in national media, you have this raging debate between folks who claim, like I said,

that data centers are drinking the ocean versus those that say no data centers are actually considerably more water efficient than, say, a golf course. To what extent do you see the environmental concerns that seem quite loud at the national level, actually moving people at the local level? I think the environmental concerns are very powerful. I think that there is something about the impacts of a data center being extremely tangible and material, which is that you have this beautiful verdant landscape, you have these lakes and trails and forests that you love so much, and now somebody is going to replace all of that with this giant,

ugly box, right? And so I think the environmental concerns, it's not even about like, oh, yeah, it's global warming or something, it's heating the planet, which it's much more like, I have a vision of my community as this like gorgeous rural place. We live by the Great Lakes, and is this thing going to threaten some of that? And so I, actually did hear environmental concerns as being almost more of a local concern about quality of life and the landscape, rather than this sort of like meta-narrative, macro-narrative about global warming a lot of the time or sort of climate change at a very big level. It's interesting because it's hard to know what to make of it because data centers environmental impacts really vary project by project, right? So for a long time, there was a ton of water that was required to build data centers because

the water would evaporate for cooling and then they would have to keep piping in new water. Whereas now, most of the new data centers are closed loop like AC. They recirculate the water so that it isn't actually using more than, say, like, the golf course next door. There are environmental concerns that I think are pretty legitimate concerned with the power that's required for these things. I mean, the Stargate Center in Celine, Michigan will require 1.4 gigawatts of energy. This is, I think, like, an additional 25% on top of DT's normal load. That's going to require new power plants to be built out, right?

And there's a question of how much of that's going to be renewable, how much of that is going to be natural gas. Some data centers are building out their own gas turbines like Colossus in Memphis. And that has very serious impacts, air pollution-wise, emissions-wise, on the community. It really depends, sort of project-by-project. And one area that I do think that, say, good policy or regulation can play a role, is in setting sort of the environmental standards for what kinds of data centers are acceptable to build to keep these to a minimum,

because I do think that's something that people are really concerned about. So we talked about AI, distrust of corporations and government, the ability of media and social media to nationalize local crises so that maybe a data center opponent in Texas can say, well, I heard what Foxconn did in Wisconsin, and this is just what corporations

due to local communities, and then you have this environmental piece. Before I move on, is there something that I missed? Is there a major ingredient to anti-Data Center politics that I missed before we talk about some of the ways that the AI community can potentially react to this? Maybe I would add one small one, which is linked to some of these other ones, but I would say affordability is a really big one, particularly in the context of electricity costs. And so both I, this was really clear in Michigan, where,

where the monopoly utility providers like DT and consumers energy are both very expensive, very unreliable and spend a lot of money on state legislative and governor's races. And so folks had been primed there by just DT has raised my rates four out of the past five years. Sure, OpenAI saying they're going to cover any rate hikes caused by the data center, but why would I believe them when I'm working with this electricity provider that's literally made electricity more expensive and less reliable every year? And then similarly, I was talking to Kathy Hokel, governor of New York about her own data center listening tour she was doing in the state.

And her team sent over some of the YouTube videos of the listening sessions that they did with local officials, with residents across New York. And more than anything, people were worried about electricity prices because these things are putting such immense strain on the grid. And when affordability is sort of, you know, the name of the word of the day in democratic circles, I do think that people's worries about electricity rates in particular, maybe deserve almost a lot of like their own sort of category. So let's say someone from the AI community jumps into this interview and says, hello, can we talk about how data centers are actually good in some ways?

Data centers create jobs for electricians, for crane operators, for earthmovers. They bring in tax revenue. Their environmental impact is often overstated by an order of magnitude. You spoke to Terry McGowan, an operating engineer, who offered this quote that I'd love your reaction to. when I was a young man, noise was money, Terry told us. Milwaukee just revered these huge, massive industrial factories because this engine is paying our bills, sending our kids to school, feeding us, housing us. A lot of those foundries had furnaces, the emissions that would come out, the slag that would come out, the oil in the machine shops, it was all pollution. But now data

centers are like public enemy number one to many of these people, and it's clean technology. So he's essentially saying, look, I remember an age, you know, 30, 40 years ago potentially, when big thing comes into town equals tax revenue, let's get the big thing into town because big thing in town equals we get nicer schools. Yeah. To what extent do you think the politics of data centers is so negative in part because, like, there is no positive coalition organized at the local level. It's like where are the Facebook groups for earthmovers, you know?

Yeah, totally. What do you make of this lacuna, this gap in a local argument or in a local case for these data centers, despite the fact that certainly in the short run, they clearly are job creators? Yeah, I mean, Terry was one of my favorite characters, one of my favorite people I met, because his perspective was, it was surprising to me. It was unusual. I knew that people working in the building trades would be relatively pro-Data Center, but he was thinking about data centers within the context of this broader economic history, right?

And so I also love that noise as money quote. I think that a lot of the stories that he told us were about the way that people's attitudes towards the economy and economic growth and industry had really changed throughout his lifetime. Of course, there's the fact that, again, these are areas whose economies were built on heavy manufacturing. This stuff is very polluting. Like, that GM plant, the contamination it left in the ground from all the fuel runoff and whatever, like a data center would not produce that level of contamination, right? And people were not nearly as skeptical of it.

Another thing that Terry was telling us was, you know, these days people are really worried about buying out farmland for data centers. And he said, when I was a kid, people hail those farmers as brilliant businessmen for managing to sell their family farms for multi-millions of dollars. People would say, oh, my God, you're so smart. And now people are saying, oh, no, but the farmers, but those farmers are laughing all the way to the bank, is what he told me.

And so he seemed kind of perplexed in general by the way that what had been seen as a boon for communities, which is economic growth, i.e. jobs, a tax revenue, i.e. everything else, was now perceived with suspicion in a way that when he was growing up, it was not. And so I found that incredibly interesting. He also told me, to your point about,

where's the local pro data center faction that he had been trying to get some of his workers to speak up in the city council meetings, right? Because the current thing that happens in most of the city council meetings is the local officials say, yes, let's do this. Tax revenue, it's good. They're saying, like, this will prevent us from having to say raise property taxes on residence. That's a big promise that's often made. The companies will give these sort of marketing slideshows that people really hate about, like, look, we're like so good for the environment. And like, look at these, like, beautiful, like, AI mockups of, like, what. this thing is going to look like. And then you have like 100 community members who are worried about

all of the things that we've talked about. And so Terry was telling his workers, hey, you guys are like, you know, more sympathetic members of the community. Why don't you guys like go to these city council meetings and talk about what this is doing for you and your ability to provide for your kids or whatever? They're making like six-figure salaries. You're getting a hundred dollar cash per DMs. It's an amazing job. And apparently people were just too scared. They were like, I'm going to get death threats for my neighbors. Like, I don't want to do that. I mean, I talked to this guy, Adam Schultz, also, who's a commercial real estate broker in Janesville. I mean, his family's from Rock County. He's been there for a long time in the community. And he was willing to speak up

at City Council in favor of the Data Center project, mainly for tax revenue reasons. And the fact that he was like, I've tried to sell this lot. Like, no one else is going to buy if it's not the data center. And he said he got flack from his next door neighbor about this project. Right. Like these towns are small, and when 70% of people are against a thing, you don't really want to be the one speaking up. And so because it has become such a polarized issue, I've heard repeatedly from people in different communities that folks who are, you know, relatively pro-data center, whether because of jobs or taxes, have actually been scared to speak up in public about these concerns. And I think it's because it's very hard to make a case for something as abstract and in some way speculative as taxes that will arrive in 10 years or jobs that are amazing now, but only for 500 people and the jobs also won't be here in 10 years. I also think that, and tell me if you disagree, there's something about what a data center is that's just different than a factory or a home. Yeah.

You know, sometimes the opposition to data centers is lumped under this concept of nimbism. And, you know, as the co-author of abundance, I'm certainly a yimbi rather than a nimbie when it comes to houses. But houses are houses for people. And data centers are basically houses for computer chips. Yeah. For which people serve as assistance in helping the computer chips do their job, which is help me answer a question that I, I've given to Claude. And so I do think that while there is an economic history to tell here that is very interesting of the growth machine of America between, let's say, 1920 to 1950,

1960, being replaced by an anti-growth machine of more zoning laws and longer permitting and more aversion to building that has characterized America in the last 50 years. I mean, this is what the beginning of our book was about. There is also this material distinction between building a house, building a factory with people operating the machines, and building that which a community understands to be a buzzing house for silicon that will not create the same kind of long-term employment benefits

as, say, a car factory. Is that a fair summary, or do you think people that you spoke to might disagree with aspects of that? I think you're right. I don't think it was something that folks talked about explicitly, but I spent a lot of time trying to puzzle through this question of, is it just nimbism? Is this different from opposition to, say, solar farms, which use very similar tactics. You have, say, farmers and rural communities,

organizing on Facebook, packing city council, whatever. And when I thought about it, I do think that's the reason that data centers are so much more unpopular than solar farms or car factories or even chip factories or things like that, is that it doesn't feel, the outputs of a data center don't feel real and tangible and useful to people. It's not even that they're evil. It's just that they're useless, right? Everyone knows we need housing,

and everyone knows we need energy, and everyone knows we need cars. Maybe they don't want it in their backyard. That's like normal nimbism. But there's a sense that, like, somewhere in your state, somewhere in your region, there should be a wind farm or there should be a power plant somewhere. As for AI, people are like,

I don't understand why we need it at all. That was a sentiment that we sometimes heard. It was just like, we don't need this that bad. There's a huge difference between, I'm a huge fan of housing. It just doesn't belong there because of the character of the neighborhood. And no, I'm against that data center, but I'm also against all the data centers. Yes, which is why you have these, like, say, way more historians. And if you look at the polling, I was talking to say Charles Franklin, who de Bruns and Marquette law polls,

and he said that one of the most interesting results is that a couple things. one is whether there's a data center proposal or site in your zip code has nothing to do with your opposition to data centers. People are just as anti-data center when it is totally far away versus when it is in their community. That feels like not nimbism. It's not about their backyard, right? And then the other thing that he mentioned was a poll of Wisconsinites where it's like, okay, of the negative impacts of data centers, which ones are the most important to you? Or like, what are you the most worried about? Number one was water use. Number two was, I don't think we should build AI, right? Which is really interesting. It was just like, why are we even doing this thing

at all? And again, I'm not saying that people think that chat GPT is fake or evil. They're just like, it's a widget. It's not that useful. The folks I know in San Francisco, maybe they feel like they can do five times as much work as they used to because of cloud code. People think that this is going to, you know, transform military advantage in cybersecurity. I don't think that story is very clear to people in a lot of communities and a lot of jobs in America, that this stuff actually does something. And so when you don't think that these giant boxes are producing something that is tangibly useful to your community or yourself, it is very hard to make the positive case. I want to talk a little bit about the national scene here. You spoke to Kathy Hochel,

governor of New York, who is very interesting on this. She is for AI, uses AI, has bragged about using AI, has bragged about welcoming chipmakers to the state, but is also announced a moratorium on data center construction. I think it's an interesting combination of policies. Texas, which many people think of as synonymous with just build the damn thing, just announced a different kind of data center pause. Opposition to data centers is rising in Virginia, which is the mecca of data center construction, builds more data centers than any state by far. What happens if America just stops building data centers? Yeah, I found some of these news pretty surprising.

I do not think of Greg Abbott or Kathy Hochel as anti-business, anti-growth politicians, right? And so, sure, like a DSA Francesca Hong, she might be running on a moratorium, but it's very surprising to me to see that policy passed in New York. And so I think that there is something serious and something. something shifting about the backlash. I'll first say, when I ask folks in the AI industry, are you worried that the public backlash is going to halt or pause the AI buildout? In general, I actually don't hear a lot of concern. It's somewhat surprising to me, but people tend to say, yeah, I think that there is a lot of local backlash, and it's very annoying,

but some states are going to welcome it, just like fracking. So it's just going to go somewhere else. Maybe the Dakotas will be really happy to have data centers. Maybe, again, Texas has temporarily now pause things, but until very recently, you know, we'll just put all the data centers in Texas. We'll put it in Louisiana. They love the data centers. They're throwing out subsidies, right? And so there's the sense that eventually the data centers would just go to the places that people want to put them. But I think that if we continue to see these pauses or moratoriums in some of the biggest states, like in Texas or if Virginia introduces something similar, we're going to have to shift more of the build out overseas. That's already happening, to be

clear. I think Australia is one place that folks are looking at to build data centers in. Some countries in Europe are thinking, like, oh, maybe actually, if we offered to build some data centers for the U.S., we can manage a compute deal where we get access to frontier cybersecurity models in return for building out AI infrastructure. Some of the things that people are more worried about is whether the buildout goes to places like the UAE. Because, again, if these things are so critical to the supply chain for American AI, you don't necessarily want to give that infrastructure supply chain choke point to an authoritarian state, right? That's pretty worrying. And so some of the concerns with the data center buildout are if they're not going to,

if we don't have enough land or we don't have enough states or communities in the U.S. that are okay with these things, we are going to go overseas. Sure, Canada, Australia, wherever at first, France, Germany will see. Each of those countries is going to have their own negotiations with their residents about whether they want data centers in their neighborhoods. And a lot of people are going to be pretty skeptical of that. But you don't want to keep going down the stack because the data center, you know, builder of last resort in the UAE is not something that is advantageous to the U.S. geopolitically from a national security perspective, right? And so the other case that happens is maybe we just don't have enough compute, I would say. The other thing that is already sort of happening

is that we have this macro trend of the closing of the AI frontier. I think this is really underrated because for the most of people are using free chatub ET, AI seems abundant. But actually right now, we're already seeing that the very best models, whether it's like Claude Mythos, for example, or even just like the more expensive chaty BT plans that can do more work at higher complexity levels, they're being a paywalled. Like you've got to pay 200 bucks a month or whatever, and we know how much money tech companies are spending on this, or restricted for national security reasons and doled out to a handpick set of, say, 50 companies and the U.S. government. that have been specially approved, right? And part of this is a safety decision. It's people saying we don't want to give a model that can, you know, produce bio weapons to a bad actor or to a terrorist.

And that's fair. But part of the reason the frontier is closing is also cost. It's that we don't have enough compute right now to serve all of the AI queries that everybody wants. This is actually really scarce. And in that world, let's ration it out so that's really expensive and only goes to a few people. So if the compute build out doesn't continue, maybe that I'll actually make AI access much more unequal.

It's going to be reserved to only those with the existing capital and ability to pay or the relationship with AI companies to say, hey, I'm like a big U.S. company. You can give it to me, but like screw all the startups, screw all the people around the world who are not going to get access. So let's say you talked to someone from Anthropic, OpenAI, Trump's AI team. And you say, look, these are the ingredients of the opposition of data centers on the ground.

they think corporations lie, they think their governments are crooked, they look at their recent history of companies coming in and promising 10,000 jobs and only creating 50. And these folks say, what should we do?

Like, what can we offer or what can we say that will change material reality on the ground in Mount Pleasant, Wisconsin? Is there anything they can do? What do you say? I've had some of these conversations,

because after I published the piece, of course, people working on the bill to have reached out, and they're like, I'm dealing with this right now. Like, everyone hates us. Like, what do we do? Like, should we just make the bribes bigger? Like, you know, like, do we just need to give people more money? I love the way they call it the bribes. I know. I think that's exactly how they communicated to the mayors too. Should we make the bribes bigger? And by the way, let's record this and send it into the local newspaper. Oh, I'm sure it works super well. But yeah, I mean, honestly, the fact that people's attitude is even can we make the bribes bigger tells you a lot. lot about some of the misperceptions of what's really going wrong with the data center backlash. But yeah, I mean, I think that one thing is that there's just kind of this misunderstanding, again, of why people don't like these things and telling them, actually, it's like a few thousand gallons of water and not 100,000 gallons of water. It doesn't really persuade people to say, trust their government. Or, again, saying like DTE is going to cover the rate hikes with Open AI. It doesn't work if people have not trusted DTE.

for the last 10 years or something like that. And so one is, I think that there is just like more understanding of what's actually going on. I'm not exactly sure what to do, but some things that are interesting are I do think that I've become more sympathetic to the state level moratoriums over this reporting. I think that if you asked me a month ago, what do I think about data center moratoriums? I would say that doesn't seem like a very effective way to regulate the AI industry. I think that there are more direct ways to do that. I think that we need to continue the build out in some form.

It seems like a very blunt, undirected tool. But when I talked to say Kathy Hochel, one thing that I became convinced of was actually because so much of the problems with these deals is about the asymmetry between these small towns, often like 3,000 people like part-time mayors, negotiating with these gigantic corporate actors and there's armies of lawyers. One thing that states can do is set sort of a floor for deals,

or streamline the standards and the terms by which a data center development developer can show up in a community. You can set at the state level regulations like it's got to be closed loop cooling. You need to use X percentage union jobs. You need to use X percent renewable energy. We're going to do rate payer protection at the state level. And sort of making sure that there is not this gigantic asymmetry between a small town alderman and these big corporations who are trying to play communities. against each other and extract as much as possible. I think that actually moving up to the state or even potentially the federal level,

what does it look like to do a fair data center deal is going to be necessary? So like at the level of like what are the terms for these things, I would say like go to the state level and like figure out what the fairest and is mistransparent way to do this is that maximizes community benefits, etc. But I do think that a lot of this is also just, it's now bigger than just like an individual town because of the sort of nationalization of the backlash, because of the way that data centers have become woven into much bigger stories about AI billionaires coming into your town, paying your mayor off to shut up and not tell you what's going on, arranging a backdoor deal, and then transforming the literal physical landscape of your community to serve some tech billionaires ego race. or whatever. That meme is big enough now that I don't know if it's ever going to be easy again. I think there's a lot that data center developers can do to do more public engagement, to do better town halls, to figure out what community benefits people actually want, to explain sort of

the job story, the tax story, whatever, to stop with NDAs and all of this. But fundamentally, I think that the data center backlash now that's become an avatar for people's broader frustrations with the economy, with tech billionaires, with the oligarchy, it's going to be a lot harder to stop. So, Jasmine, last few questions I want to turn the table a little bit. I think of your work as fitting into this genre of cultural translation. You translate America for people who live in the Bay Area and build AI, but you also translate or interpret the culture of the people who are building AI

for readers in the rest of America. And I wonder, what does that? the one or two ways that you think Americans would be most surprised by the philosophy of or the personality of the people who are building this technology? I'll say two things. The first one is that I think that many people building AI from the researchers to the executives to the CEOs of these labs are in fact deeply unsure about whether the things, thing they are building is going to be good for the world. I think that their concerns about cyber attack risks, about job loss, about even concentration of power and capital and their own

companies, I don't think that these are a marketing hype. I think that these folks genuinely believe them. They're worried about them and they feel caught in a race, a capitalistic race, a geopolitical race with China that feels bigger than themselves and that they have no idea how to stop. They're all racing forward because there's a sense that somebody's going to build super intelligence. Maybe I can build it the least bad way and also get rich, but I can't stop it. And so when they say things like, we wish the government would regulate us more, I actually think that's true. I think that a lot of folks are scared by the lack of agency they feel and being able to stop their own creation. And so things like this slowdown letter that has been signed

by all these executives and at these big companies saying, we need to figure out a way that if we wanted, we could pause the development of AI. I think we should take those desires pretty seriously. And then, of course, it's the government and policymakers' jobs to actually figure out the regulation. And that's going to be its own battle. But the extent to which people in the AI industry often feel very conflicted about their own project, I think is worth taking seriously. The second thing that is important to understand about AI that follows from that is the number one question I get from folks outside SF outside AI is if they feel that way, why are they building that? If I was worried that the thing I was building had a 10%

risk of blowing up the world or taking everybody's job, I just wouldn't do it. That feels like way too high risk for me. And there, I think there's something more important to understand about Silicon Valley's attitude towards risk, which is really unusual. It's a community that prides itself on being risk takers, on I'm going to, you know, invest in 10 startups, and it's okay if nine of them fail. It's okay if I bankrupt myself on the way to trying to achieve. glory and immortality and whatever else. It's okay if I inject these Chinese peptides that might be full of substances I don't understand, but might also make me tanner and hotter and healthier, right? And this is a community where people are willing to take these sort of high expected value bets,

where, you know, there's a famous quote from Sam Bankman-Fried on a Tyler Cowan podcast where he says, if there's a 51-49 coin flip, 51% odds, I could double the amount of human welfare in the universe, 49% odds, everyone dies. Sam Beckman-Fried says he would flip the coin. I think that's insane. I think that's ridiculous. And at the same time, while that example is very unsympathetic, I think that for a lot of folks building super intelligence,

maybe it's not a 51-49 set of odds, but it's something like 90-10, and they are still willing to flip the coin in that world, right? And so the way that folks' attitude towards risk and towards I am willing to take the gamble at immortality and super abundance and AI that loves us and benefits us all is something that I think is very different

from the type of gamble that a lot of folks elsewhere in communities being materially impacted by AI are willing to make. I'm so glad that we ended there because you sparked something that I had never really thought about before. We've used this term power asymmetry

again and again in this show, I think appropriately, because of the power asymmetry that people in Wisconsin and Michigan feel about these large companies, trillion-dollar valuations, billion-dollar CEOs, making deals with, cutting bribes with their democratically elected leaders in order to put this enormous, disgusting, ugly building situated

in what is otherwise an Eden. There's a power asymmetry there. But what you're describing is subtly different, but also really important. It's not a power asymmetry. It's a risk asymmetry. You're talking about a Silicon Valley culture that prides itself on risk that says, you know what? quadruple leverage yourself on Korean stocks and, you know, wipe out your entire fund.

It's okay. You'll raise money again next week. Invest in 10 companies, fail in all of them. That's fine. Spending your entire life building a technology that has a double-digit percent chance of wiping out the human race. Well, there's a 90 percent chance we get to super abundance, so go ahead and do it. You have this extraordinary appetite for risk, which is attached to money in Silicon Valley.

And you juxtapose that with what I'm thinking back across our conversation. as a very different attitude toward risk among these people who live where data centers are being situated, they're thinking if there's a 30% chance that this is another Foxcon situation, then hell no. If there's a 20% chance that this data center or AI breaks down the same way that the GM factory and America's auto industry broke down, then hell no. If there's even a small chance that they might be lying about some of these numbers, then hell no. that if there's any risk, then the answer is no.

And that's a really interesting juxtaposition. It speaks to, I want you to have the last word here, but I'm just thinking it's just really interesting because it speaks to not only this economic imbalance between America and Silicon Valley, but also this like characterological or psychological gap between the people building this technology and the people living around the data centers

they're plotting across the country. Jasmine, last word here, please. Yeah, I just want to emphasize that I both think that there is this deep values and dispositional difference that comes with building your career in a high-risk industry like startups in VC and AI. And I think that a lot of Americans correctly understand that when you are playing these high-risk games, who gets left holding the bag is not an equal endeavor. The distribution of risk is not an equal endeavor. Leopold Oshambrunner can blow up his fund and he will be able to raise billions of dollars again. Whether when that GM plant

goes away and 7,000 workers are laid off overnight and they can't retrain and they can't move, that is a very different distribution of risk. And I think that a lot of folks kind of correctly suspect that if this is a bubble, if the AI thing does blow up, these AI people are going to be fine. They've got enough money over there. They've got the next tech project to work on. But the people in these small towns with a giant data center of the community that the tax base has become dependent on now, they are not going to be able to recover as easily from

that. Fantastic point. Jasmine Son, thank you so much. Thanks so much for having me.