AI is eating the American economy. But American politics is eating AI.
Very often, the most important thing you can do as a writer trying to see the world clearly is to “stack facts.” That is, don’t worry too much about conjuring grand theories or lurid forecasts. Just write down the things that you know are true. Put them next to each other, like dominoes in a line. Once you’ve done that, you don’t have to do anything special. Pay attention. Look for breaks in the chain of logic. If the facts point somewhere that doesn’t make sense—if the line of falling dominoes leads to an end point that is strange, or surprising, or even catastrophic—then you know you’re looking at something interesting.
So, for example, here are three indisputable facts about the economics and politics of AI.
1. Whether it thrills or terrifies you, artificial intelligence has never been more important to the US economy than it is right now.
Start with overall economic growth. AI spending currently accounts for about one-third of economic growth. If that doesn’t mean much to you, some historical comparisons might help. The AI boom is sometimes compared to the dot-com bubble. But the data-center buildout has actually significantly overtaken the dot-com boom, as a share of GDP. In fact, AI spending is now rising faster than home-construction spending during the boom (bubble?) of the mid-2000s before the Great Recession, according to Torsten Slok, chief economist at Apollo. This is the biggest and fastest-growing infrastructure project of our lifetimes.
Or take stocks. The best-performing stocks of 2026 are almost entirely tech firms building the physical infrastructure of AI, including storage and memory firms, chip and chip-equipment manufacturers, and networking and optics firms.
Despite the double albatross of Trump’s tariffs and wars, the U.S. continues to chug along, in large part because of the historic contributions of the AI boom. On August 3, the *Wall Street Journal *put an appropriately hyperbolic headline on this whole situation: “The AI Boom Is Transforming the American Economy Beyond Recognition.”
2. The AI boom desperately needs to build more data centers.
Data centers are big buildings filled with chips, memory, storage, and networking equipment required to train AI models and to run them for users. The latter process is technically known as “inference.”
Consumer demand for inference is surging for four interlocking reasons. First, more people are using AI. Second, the typical user is asking more questions. Third, the typical question is eating up more tokens, especially as coders make use of autonomous “agents.” And fourth, powerful new models and agents often require more computation per token.
The result: a staggering rise in token consumption. Nvidia CEO Jensen Huang said last year that “AI inference token generation has surged tenfold in just one year,” and Google’s monthly token processing increased roughly 300-fold between May 2024 and May 2026.
All of this work has to happen within data centers. So, it is not hyperbolic to claim that the most important industry in the US economy is dependent on the ability to build more data centers to train and serve inference to its customers.
3. Opposing AI—and, in particular, opposing the construction of new American data centers—has never been more popular.
Conservative Republicans and Democrats don’t agree about much. But both groups of voters really don’t like data centers. In fact, their antipathy to data centers might be the only thing they strongly agree on these days.
“I’m not sure I’ve ever seen a chart where conservative Republicans are closer to liberal Democrats than liberal [and] moderate Republicans are,” Anthony Leiserowitz, the director of the Yale Program on Climate Change Communication, has said of the following graph.
Overall, negative attitudes toward AI are higher today than any previous period of polling. Noah Smith points out that data centers are more unpopular today than nuclear power plants ever were in the history of polling; and the U.S. practically stopped building those for decades.
The unpopularity of data centers is having a significant effect on the ability to build them. Hundreds of groups around the country are coordinating to block the construction of data centers. Despite New York Gov. Kathy Hochul being a proud user of AI and a champion of chip manufacturing, she declared a data center moratorium in the state. Despite being known as the Mecca of permissive construction rules, Texas recently announced its own data center .
To review: * The fastest-growing political movement in America is dead-set on stopping the construction of data centers, which are critical to the growth of the fastest-growing industry in America.* This is not an interpretation or an opinion. It’s just the stacked facts. American politics and economics are on a collision course.
So what is the data center backlash really about? Is it based on fear and viral myths about data centers sucking the oceans dry, or is it based on facts and local knowledge of faraway corporations lying to small towns? Jasmine Sun is a writer who recently went on a reporting trip to the Midwest to find out what’s really behind the national AI revolt. In today’s conversation, we talk about how opposition to AI became the biggest applause line in politics and and what happens if the anti-AI forces succeed in shutting down the data-center boom.
Derek Thompson: You live in San Francisco. You know the people who work on artificial intelligence. What do people in the Bay Area think data center opposition is all about, and to what extent did your reporting validate or invalidate those assumptions?
Jasmine Sun: In San Francisco, the default caricature of a data center opponent is someone who’s watched too much short-form video and absorbed myths like data centers sucking up the oceans or emitting radioactive waves. The water meme in particular—the idea that AI data centers use way too much water, which the research doesn’t really support—has become so prevalent on social media that it’s given a lot of Silicon Valley people the impression that opponents are simply misinformed.
But on the ground, opposition was much more about residents losing trust in local officials and in the way big corporations interacted with their land. I almost never heard people talk about OpenAI or cite specific AI harms, such as job loss or deepfakes. What I heard much more often was: How could my mayor sign this NDA and keep the project secret from us? How could my city council sell out to this giant corporation with no public input? Fury at the lack of transparency and the sense that backroom deals were corrupting small-town democracy was much more the story on the ground.
Why America Hates Data Centers, Reason One: Is It Just AI?
Thompson: Your reporting showed that opposition to data centers wasn’t about any one thing. It was several themes or motivations. I want to go through each of them.
First, let’s start with AI. Before reading your piece, I would have assumed the opposition to data centers was rooted in Americans just not liking AI. How much of what you found was about that?
Sun: Almost none of it.
Some people mentioned AI as a kind of cherry on top: We’re building all this for a technology I don’t even like that much. But the activists showing up at city council meetings weren’t focused on the risks AI posed. We almost never heard people say the words ChatGPT, OpenAI, or Anthropic in these hours-long conversations.
That said, the fact that people didn’t like AI much made the trade-off feel not worth it. People were more open to new development when they could see societal benefits, the way they might with a solar farm or a car plant, and AI seemed less useful than either. With the exception of some more ideologically motivated activists, most people were concerned with material impacts, government transparency, and general distrust of corporations rather than the specifics of what was being built inside these buildings.
**The Biggest Source of Anti-Data Center Politics: An Explosion of Citizen Distrust **
Thompson: So narrow antipathy toward AI played a smaller role than expected. Did broad distrust of companies and government play a surprisingly large one?
Sun: This was one of the most fascinating things to me…