Is This the Fastest Opinion Shift in American Politics?

I'm David Wallace Wells. I'm a writer for New York Times opinion and a columnist for the Times magazine. This is an insanely polarized country, which means it isn't very often you see a political supermajority on any issue. Let alone one that emerges almost out of nowhere to announce itself as a major new front in our politics. But that's exactly what has happened just over the last year with AI data centers. Around seven in ten Americans now oppose the construction of local data centers to power AI. Last fall, people were almost exactly evenly split when asked if they'd support a new data center nearby. Now, there's a 50 point gap. That is an absolutely crazy swing. Much bigger than the reversals you tend to see when a new president comes into office or even when someone declares war. So what's actually going on here? Is there suddenly a new populist politics sweeping the country? Can it be channeled into something productive? And does the political economy around big tech and AI even allow for something like the democratic control that people protesting data centers seem to want? I don't know where all of this is heading, but I do think it's really significant. When I wanted to talk this through with a friend of mine, Robinson Meyer. Rob is an opinion contributor and the founding executive editor of the great environmental news website heat map, which has also been doing a ton of important work following the data center backlash. Rob, thanks for being here. Thank you so much for having me. It's it's great to be here. So we're going to talk about the politics of the data center backlash, which are really interesting, but let's just start with the sense of the scale and just how fast it's come together. Have you been as shocked as I am about what we've seen over the last year? It's been so shocking because I think it's the it's a scale of change in public opinion that I didn't know was still possible in the US. I think you mentioned it in the intro, but I think we're used to a new president coming in and everyone, if it's a Democrat getting a little more conservative, you know, if it's a Republican voters getting a little more liberal, but we're not used to seeing these huge issues pop up on which the public goes from having one view or probably not really having a view to being very hardened in their opposition. And that's what we've seen with AI data centers. So to throw some numbers at this. Yeah, please. So a heat map, we cover climate change decarbonization and the energy transition because of that we started covering public attitudes toward at first renewables. And then because of the scale of the AI boom, we started asking about AI data centers as well. And so every quarter, we pull about 4,000 Americans and we said, would you support or oppose a data center being built near where you live. We asked that question in August of last year and Americans were very split. They're about 43 in support and 42% against. We asked it again in the spring. And this time, 21% of Americans said they would support a data center being built near where they live and 71% said they would oppose it. So it is a staggering shift basically in how Americans feel about this industry and about this form of land use in their communities. And let's walk through what that means in some more particular ways. I mean, this is a movement that draws on a lot of threads. You also see different intensity. Some people really, you know, eager to go out to a town hall and protest the data center, other people just telling someone is taking a survey that they're not sure that they want something being built down the road. So beyond the top line numbers, what else are you seeing over this relatively short timeline? Yeah, totally. So at a heat map, we also track. Data center opposition. So basically, how many data center projects across the country are being opposed by their local communities and then how many are being canceled. But what's interesting there is that, first of all, the numbers have skyrocketed. Second of all, you tend to see some interest like an interesting landscape for why people oppose these things or what comes up. You know, in our data, the number one reason that people oppose data centers in their community is water. They're concerned about water use. The number two reason is electricity. What's interesting is that those are not actually when a data center project actually gets canceled. Those aren't why it tends to get canceled. Why it tends to get canceled when it comes up as noise. But water and electricity are very unifying concerns. And what's interesting is like among older Americans who are opposed to data centers and among younger Americans who are opposed to data centers, they kind of all agree on those elements. Among younger Americans who are opposed to data centers, you also see suspicion of AI and a mistrust of AI and a dislike of AI driving reasons why they might oppose the data center in their area. It's also the case right that older Americans tend to be somewhat more supportive and younger Americans, somewhat more skeptical. Right. I think what's interesting is that if when you ask people to like what are compelling reasons to support a data center, like Americans understand all the benefits as they see at the data centers bring to their life. They like that data centers enable streaming. They like that they provide local tax revenue. They like that they create construction jobs. What's interesting is that then they don't like as much when they are described as enabling AI. And so it does seem to be AI data centers specifically that are picking up the backlash. So I wanted to unpack this backlash ideologically. And you know walk through at least as I see it, the sort of comprehensive list of objections, which aren't exactly contradictory, but are also not exactly always, you know, they don't line up. In the most intuitive ways are draw all the same people together in unison. You know, you mentioned water objections. There's also the effect on electricity prices, the burden on the local grid, local pollution from any extra power generation that's required. There's also noise. And that's all sort of like in one bundle, which is to say the local impacts of this construction. And then there are objections to the purpose that this is being put towards and to that. And I would say, you know, anxieties about AI and what it means for the future of the economy, the future of employment, to some extent, the future of our social lives and our political lives, any even existential anxiety to the extent that that's playing a role. There's also worry that people have about the particular political economy of this business. And the fact that we have a few companies run by a couple of people who are insanely wealthy and through AI seem to be laying claim to more and more control over the future of the country and the economy. People look at that development pattern and feel so powerless. And they think, you know, if Sam Altman is going to decide that my community is going to suffer water shortages and local pollution and my electricity bills are going to go up and me and everybody I know is going to be put out of work by this. And I have no lever to pull my access to power in Washington feels really limited. What can I do here is a town hall I can go to and scream my ass off and get this project killed. And so there's, you know, there's there's that piece of it too, but wonder if we could talk about each of these both as sort of concrete claims that is like how serious are the problems about water and pollution, for instance, and also what it means politically to be mounting a political objection on this ground. So, you know, my read of the research is that the environmental arguments. And those claims are often a little overstated. There is a effect on water. In some cases, it can drive electricity prices up, although in other cases, it drives them down. But it is interesting to me that the this set of claims, which are the center of the political energy there are kind of most factually dubious. Is that the research there? It can be hard to talk about this because every project is different. And so I hesitate to say, well, this claim is overstated because you can point to this particular project or project or another particular project. I would say that my reading of the research here as well is that water tends to be where there are the most let's say like folk claims about data centers floating around that are not true. But people have this idea that every time they use Chatchee PT or Google something now, a whole water bottle is being poured out of, you know, from some precious ecosystem that probably isn't true. And in fact, the tech companies are very attentive to the water concerns in part because they are so unifying. I mean, I think one thing we see in our data is that Americans across political parties across age cohorts agree that water and power prices are a problem. And then we get into these questions of electricity and potentially also local air pollution to some degree, those questions are linked. I mean, I think what we haven't yet seen in most of the country is data centers driving up power prices with one exception, which isn't the mid Atlantic in the electricity market that covers New Jersey and Pennsylvania and Washington DC and all goes all the way out to Chicago. We actually do have pretty good evidence that local data centers have driven up power prices and that's partially because of how that electricity market is designed. But outside of that, we actually don't really see yet evidence in the past that data centers have driven up power prices, what we do see, believe will happen is that they will increase in the future as data centers basically bid for what could be quite scarce power over the next four years. I think where data centers can be said to have a negative influence on the environment more broadly is on climate change and that's just because they're a big electricity user and we haven't decarbonized the electricity system yet. And I think what's perhaps most troubling is that any world in which we decarbonize, we're going to be using far more electricity than we use for data centers today. We're going to be using electricity to run our cars, we're going to be using electricity to heat and cool our homes. I deal for everything. Exactly. And that's all going to require far more electricity generation than we have today. But we're struggling to meet the generation needs of data centers, but we have to build so much more generation than we have right now just to meet the needs of our economy in a decarbonized world. You and I talking about these issues, it can sound extremely technocratic. Like we're designing this whole system in this little room, we're working on a whiteboard, we can see the trade-offs and we can then manage the trade-offs. It's almost like we're making a case for the AI build out and that everything is going to be okay. But of course, the energy on the ground is not technocratic, techno-optimist, people in a room crunching the math. The energy on the ground is at least as it seems to me like pretty heated and that heat is pretty widespread also. And I think that that goes beyond particular objections to water use or even local pollution, there's something like deeper and larger that seems to be unfolding here. And it puts me in mind of, you know, occupy, it puts me in mind at the tea party. You know, when you see videos of these town halls, people are pretty angry and one thing I've been struggling with is trying to figure out just how thick that opposition is. We can see clearly in poll after poll that more people are opposed to local construction of data centers. We also see people are worried about AI more generally. But to the extent that we're then thinking about how this plays out going forward in the midterms and the presidential cycle at the national level, I would love to have a clearer sense of how robust that opposition really is and what it points to when we think about, you know, the possibility of taking control or taking hold of this sort of AI industrial revolution that we're living through. It's funny because I, having looked at the data, I am less convinced of the thickness of this trend. I think, first of all, this trend so far has mostly not congealed into something on the scale of like an occupy. People are opposed to data centers in their communities. They're beginning to talk to each other. But I guess where I, here's the challenge that any data center, any widespread data center movement would have to meet, which is that once you get past local projects, there is not as much agreement about why people are opposed to these things, kind of in an ideological or national way. And the coalitions that spring up around projects, or at least the coalitions that we see in polling, when you look under the hood, what you see is like, okay, there's a lot of agreement about local electricity and water use. There's less agreement about what we term in the data environmental impact. What we see is that environmental impact of data centers, which kind of covers water and air conventional pollution, is like a top concern for Democrats and younger Americans, and not a top concern for older Americans. So like, some are all that what you mean is like that this has a very strong NIMBY core. NIMBY is so pejorative, but like I do think is that people who would object less if it was, if the proposed data centers were somewhere else far away from there. I think that's right, yeah, and I don't see, I mean what we do see is that most people, number one, are pessimistic about artificial intelligence and people who strongly oppose data centers are pessimistic about artificial intelligence. And that is true broadly, and perhaps from that, a movement could kind of transfer energy around opposing local data centers into opposing AI or regulating AI in some enduring way. But you start to lose parts of the coalition. You start to lose parts of these impressive numbers if you do that, let's say, you know, what's interesting is that at this point data centers, you know, net support, would you support a local data center in your community is underwater among all parties, Republicans and Dependents Democrats. If we ask, well, what if it was powered only by renewable energy? Then Democrats are more likely to like it and Republicans are more likely to oppose it. And so all the old divisions of Americans politics are still there embedded in the movement to oppose data centers. I think the other thing that's worth adding into this mix is like data centers are large, often land intensive, extremely energy intensive facilities. When you talk about what uses that much energy with the acreage of a data center, you're talking about things like aluminum smelters. You know, you're not even talking about like, yeah, exactly. Nobody wants an aluminum, there's very little support for like, yes, I want an aluminum smelter in my ex-servering community. And telling people that those aluminum smelters, we're going to put them out of work. We're going to maybe kill everyone, we're like, not really necessarily safe yet, but we have to do it for national security or for who knows why, who knows why, yeah, exactly. And on top of that, it's not even like a manufacturing build out or building out warehouses where you can then turn around and promise people that like, well, yeah, we're going to be drawing 800 megawatts of power in your backyard. But everyone in your community is going to have a job here and it's going to be an engine of economic development in your region for the next 40 years. We're saying, yeah, this thing is going to go in your backyard. It's going to have the equivalent of a small natural gas power plant and it's going to maybe be part of this larger economic trend that puts you out of a job and it's going to employ like maybe 20 people over the course of its lifetime. And so it's just not a very charismatic form of economic development. I think it's part of it. Although, you know, one of the most astonishing facts that I read about this recent pattern is that Latin County, Virginia, which is the epicenter, expects to get half of its local tax revenue from AI data centers by 2030 or so. So you're talking about a significant expansion of local tax revenue. But I think one part of the AI story that feels so distressing to people is that a lot of people will say at these town halls, nobody asked for this. I wasn't screaming about how I needed a chatbot to run my travel itinerary or do my taxes. Maybe I'll find a reason to do that. Maybe it will even be a little bit better. But we're living through this revolution, I'm speaking in the voice of someone, an activist, living through this revolution that none of us were agitating for, because a select group of very rich and powerful people decided that it was absolutely necessary, persuaded the powers that be to take their hands off of the system and let the money do its work. And I think one of the, personally, one of the things that I think is animating the local backlash is something that's not all that local, which is just the sense of my future is now being written by people who don't have my interested heart, who are completely insulated from the kinds of effects that they may be unleashing and who have bought off the at least tacit support of the federal government. And that powerlessness is quite enraging and it does make you think, where can I grab a hold of this story? Where can I take control? And I think that's one reason why it's showing up so much more at the local level, the federal level is because you can just storm into a town hall and you can just rally your neighbors against something in a way that you can't necessarily make sure your senator is going to vote for a, I did a centimortarium or whatever, there is some access point that is still visible to people in local communities that even if it isn't going to change the whole shape of the AI future, it allows them to feel empowered in a way that almost any other political activism towards AI is going to feel effectively like they're engaging in some kind of fatalistic performance. I think that's right. Well, I mean, I think that certainly like, think global act local has been true of many, many movements. And I think that that is definitely something that is driving part of the backlash to them. And I think there is a sense in all of this where the American public believed that the last several computing revolutions had really improved their life and had really allowed them to have a deeper existence and had made them clearly happier than they were 10 or 15 years before it happened, then we probably wouldn't be seeing a backlash on the scale that we're seeing now. And it isn't interesting that the international level, the US is actually considerably more skeptical of AI and worried about AI than almost any other country for reasons maybe having to do with the way that Silicon Valley has re-scripted our lives over the last couple of decades and maybe also to some extent because we've seen the factlessness of our politics in trying to take control of that system. I think it's almost, to some degree, when you see how politicians are thinking about Silicon Valley, the tech industry, it's almost like they're thinking about how to manage a resource curse. Yeah. Right. This is this huge, huge factor in the American economy and American daily life and American politics. And we know it's driving a huge amount of the country's wealth. We know it drives a huge amount of the country's competitiveness. We know that it increasingly drives a larger and larger portion of ordinary people's wealth because it's come to dominate the stock market so much. And yet, it doesn't feel like we really know how to get off the car. Where I want to be careful is that I think this set of sensations that we're describing. When you look at the data, they leap out for younger Americans, they leap out for Democrats. If you look at the full scale of people who are opposed AI data centers, I do think older Americans, you don't see it. You see more NIMBY, classic NIMBY threats, but there's no way this happens without a mass public feeling that the past 20 years or 25 years of computing innovations have somehow left us with a shallow world, a less satisfying world, a world where local communities have less control over their own lives and where there's less saver, I don't know in the American experience. Let's move to the national politics part of it. And think a little bit and let's talk a little bit about what we're seeing at the national stage. There is more rhetorical lip service being paid to the risks of AI and to the problems of quote-unquote problems of AI data centers, but it still feels to me somewhat on the margins and not yet co-hearing into something that looks like a viable policy agenda. Is that how you see it? I mean, I think in our polling, what we see is that there's national majorities among parties for a national data center, moratorium, a state data center, moratorium, and a local data center, moratorium. And yet it is not a mainstream policy. The one thing I want to kind of say here is that a movement doesn't have to congeal as a movement to be good for a political party. And I think we've already seen this movement be good for Democrats. I mean, I think Democrats are already running against AI or against data centers or against a sense that under Donald Trump, the AI industry and the tech industry has been allowed to shape our lives. And I think what we've already seen is that when voters feel like there's some industry or sector of the economy that's out of control, they reach for Democrats. They believe that Democrats are the party that can help them. And I think that we're already seeing that in midterm. Yeah, I think that's absolutely right. The sort of political story here makes sense for an out of power democratic party. Yeah, exactly. But I also wonder where that leads. I mean, if we imagine some amount of AI backlash, at least helping power, say democratic majority or majority is starting in 2027, what policy flows from that? I don't feel like there's a clear policy menu on the table for these things. I think what's interesting is that when you look at a politician like Bernie Sanders, he's put forward a national data center moratorium, other politicians have not quite followed. But when you kind of look at how Bernie is expressing why he thinks the moratorium is the right choice, it actually has far more to do with the existential concerns around AI and the Doomsday concerns around AI than it does with the local land use concerns. It's also a little bit contradicted by the fact that he's also proposed taking an ownership stake. Yes, exactly. I mean, he seems clear that there's energy here and it seems almost like he's trying out different things to see what sticks. Yeah, and I think that's where a lot of candidates are right now. I mean, I think if you take a step back, I guess what I'd observe broadly is that most of the workers in the tech industry and some of the executives were seen as progressives for the past 10 to 30 years, and the democratic party saw part of its mission as empowering this new sector of the economy in part to combat the old industrial forces that it believed were holding back the economy or didn't put environmental causes or labor causes at the center of their existence right there has been a long-term policy goal of empowering the technology industry as almost an economic development policy for the United States in part because the technology industry was going to solve problems that we wanted to see solved or the Democrats wanted to see solved. Or the very least for just providing so much more growth than any of us. Exactly. Yeah, yeah. And what changed during this Trump administration is that the tech industry identified itself with the Trump administration. It saw the Trump administration as being key to getting the outcomes that it wanted. All those CEOs went to his inauguration. And now that's partially because executives in the technology industry were annoyed or frustrated or hated policies adopted by the Biden administration, but the connection of the Trump administration and the technology industry. And I think the Republican Party with the technology industry is like a new thing that Democrats in power have never fully had to reckon with. And I think there's still some disagreement within the party about whether they heighten the contradictions further and try to regulate the technology industry and the AI industry as aggressively as for instance, we regulate say the car industry or the power sector or heavy industry or whether they try to seek some kind of compromise or return to the status quo. It's a hail lot of technology CEOs are very important Democratic Party donors. They're very important contributors to all these other causes that we care about. And we are going to try to seek some kind of comportment. We're going to seek some kind of compromise or deal with them instead of heightening the contradictions and cracking down on them further. You know, one of the things that I've thought a lot about over the last few months is the Pope's Encyclical. And one of the things that I read in the immediate aftermath of that was just a piece in compact magazine, I think, where it used as a sort of comparison point, the movement to ban or at least sanction the practice of cloning in the 1990s. And the role that the Catholic church and the Pope had played in organizing a global coalition around that prohibition. It also reminded me that in the George W. Bush years there was a movement to ban stem cell research. It never got to the level of a federal total federal ban, but there was a ban on federal funding that really changed how much of that kind of research was possible. And it still felt workable that we could in civil society and in our politics take control of the force of progress and history represented by science and technology. And you know, it didn't seem outrageous. And I just wonder whether we have any possibility like that today or whether the tech companies in particular have just grown so rich and so powerful that they seem to operate beyond the reach of conventional politics in a way that not doesn't just despair at the average voter, but makes it hard to imagine anyone to the right of Bernie Sanders pulling together a meaningful political response. I think a few things. I mean, I think to some degree what we're picking up on is not a lost moment where we believe that science and technology were governable, though that is part of the makeup. I think the other thing we're picking up on is the transformation of the Republican Party from a party that represents religious values and has a sense of clearly evangelical Christian values and virtues around life and what life should be and the role of men and women to a Republican party now that at least as embodied by Donald Trump doesn't have many virtues or accept those around power and doesn't stand for much beyond, you know, self-aggrandizement accumulation and power. So part of what we're picking up on I think is just the transition from a Christian right to a post-Christian right. But I also think that there's two other factors here. One of which is that it was particularly easy to govern global science and technological development in the 1990s because only really the US in Europe were at the forefront of global science and technological development. And so if the US in Europe agreed on something and they were allies, then everyone else wanted to work with them. I'm vastly simplifying obviously. I think what makes this moment different is that I mean we're talking about Silicon Valley demanding no restrictions on its growth, often demanding government support for whatever its leaders want to do. That is made in direct comparison to China, right. I mean, and the fact that the US is now under competitive pressure, I think is driving a lot of the elite decisions here and I think also makes any kind of global governance of these technologies. I don't know both easier and harder to imagine. Easier in that if we are going to a bipolar world where it's the US in China have frontier labs and frontier AI capabilities and if those two governments agree on reigning in their technological development, their innovation, their work on AI in certain ways, I think that actually does enable governance and I can see a future US administration regulating AI or pursuing some kind of bilateral agreement with China or trilateral with China and the EU, reigning in AI in some ways. I think when we talk about a national movement to reign in Silicon Valley or to reign in the tech industry or to convert the NIMBY energy around AI into something more enduring, something more like the environmental movement where I'm skeptical is not actually around the public's appetite for this but around the repeated failure of movements in recent history. to build anything enduring on that scale. We've seen many, many moments over the past 15 years, climate change, I would say, is one of them where it seems like there is building public anger, even cross-partisan anger, or building public support for a certain policy outcome. And for whatever reason, because publics are fickle, because people get distracted and move on. It's just because I think there's something about this moment in media, in public attention, in civics, that makes it very hard to create sustained public focus on problems in a way that builds the level of the environmental movement or builds the level of the civil rights movement. And that makes me doubtful that this will turn into something as enduring. So one last question for you, a little more pointed. If the main thing that it seems to be on the table policy-wise in the U.S. is a data center moratorium. Do you think that's a good idea? If that's the main thing on the table, no. What I would like to see is closer to a policy where we take the immense amount of investment that companies want to do in data centers, we then point that investment, maybe with an excise tax, maybe with some kind of fund, we point that investment toward making the upgrades that our economy needs anyway. And what our economy needs is a revitalized and much larger power grid. What our economy needs is ways to build out transformers, ways to build out wires, ways to build out the physical goods of the economy. All of that feels way more fruitful to me and way more likely to improve people's lives than just saying stop to this one form of technology. We'll leave it there, Rob Meyer. Thanks so much for talking. Thanks so much for having me. [MUSIC]