Iam possessed by a fundamental conviction that geopolitical supremacy is built on cutting-edge industry. And I’ve been waiting for this moment, the moment when tech companies would cease being “agile” and built to pivot and start digging in the ground with major capital-intensive investments in hardware. Data centers aren’t the main thing in geopolitics, but just as Ford’s factory lines were put to use in World War II, so it may be that these data centers will be powering drone swarms and cyber warfare in the future.

And yet, despite spending like crazy in a way that is keeping the American economy humming and employing lots of blue-collar workers, an NBC News poll earlier this year found a net-favorability rating of “minus 20” for AI. Fox News polls have found a consistently growing “concern” about AI over the past year, with the sharpest increases among women. A Harvard Caps/Harris poll in July found that two-thirds of voters want to slow down AI development.

But the spending is only accelerating. Amazon, Microsoft, Google, Meta, and Oracle are spending $600 billion on infrastructure in 2026, which is up 36 percent from last year, which itself was a record year. Tech companies issued more than $428 billion in bonds in 2025 and may yet borrow trillions more. Goldman Sachs is predicting that $7.6 trillion in capital will be deployed in the next five years across computing, data centers, and power.

This build-out is rippling across the economy, pulling HVAC and other tradesmen into electrical, which is seeing a massive demand spike, driving up salaries for senior electricians deep into six-figure territory. Mike Rowe, host of Dirty Jobs and advocate of blue-collar work, says he knows electricians who are making over a quarter million a year. BlackRock predicts that America is going to run out of electricians needed to build data centers. And we haven’t even talked about the data centers that Elon Musk is promising to build and then launch into space, where he plans to use solar power and space’s infinite heat-dump to reduce compute costs.

And yet this build-out is receiving massive backlash from the public and from citizens organizing in their localities. People are campaigning to rule out or impede the building of data centers in their communities forever. In some cases, they cite surges in electricity costs because of power-hungry behemoths moving in where there is a finite supply. A typical hyperscale data center uses about 100 megawatt-hours, as much as 100,000 homes. States with room to grow on the grid, like North Dakota, have found that data-center-building lowers costs for customers overall. But in the mid-Atlantic, the demand surge can lead to price increases.

Sam Altman, CEO of OpenAI, worried that public absorption of AI “does feel surprisingly slow.” At a tech conference this February, Jensen Huang, the CEO of Nvidia, the chip producer that is minting money off the boom, called the public backlash “extremely hurtful.”

But the AI industry has itself to blame for these attitudes. How can Altman be surprised at public reluctance when he signed a letter in 2023 saying that AI posed a “risk of extinction” to humanity? Huang, just last December, said the technology would touch “every job” and seemed to echo an industry-led narrative of social catastrophe. Anthropic CEO Dario Amodei said in May last year that AI could cause a “dramatic spike in unemployment in the very near future.” And Geoffrey Hinton, the “godfather of AI,” was reported to have said that it seems very likely “there will be unemployment,” with tens of millions of jobs displaced.

No industry in the history of the world has marketed itself with such imbecilic self-sabotage: What we’re building is horrible, will ruin your livelihood, and possibly enslave you! It’s also inevitable. Oh, and may we please have a few acres to scale up our operations?

They are doing this at a time when people are more ambivalent than ever about the social transformations that the internet, smartphones, and social media have triggered. There is, in many people, a growing unease with the costs these technologies seem to have imposed on us and a nostalgia for a technological landscape that didn’t immediately strike anyone as dystopian.

It doesn’t have to be this way. AI companies could emphasize the good they are doing, even now. They’re going to make access to ancient texts possible for scholars for the first time. They are rapidly improving medical diagnosis based on scans and pointing pharmaceuticals toward new drug discoveries. They are revolutionizing language-learning and translation. We are solving centuries-old math problems at a real clip. And yeah, as a side effect, we’re increasing the fortunes of hardworking Americans.