Let’s say a big tech company announces plans to build an AI data center just a few miles from your home.
On Wall Street, that’s incredible news.
More construction leads to more computer chips, and another chance to cash in on the relentless AI boom. For perspective, according to S&P Dow Jones Indices, the Magnificent Seven (mostly AI stocks) accounted for a head-turning 55% of the S&P 500’s total returns between 2023 and 2025.
But if you’re the one living nearby, your questions are likely to be a lot different. Will all those servers push up electricity costs? Will the project strain local roads and water supplies? Moreover, if AI is supposed to make businesses more efficient, what happens to the jobs people already have?
It’s not tough to see why that enthusiasm might look different from the other side of the fence.
That disconnect just caught the attention of Dario Perkins, a senior economist at TS Lombard. While investors butt heads over AI CapEx, Perkins sees another potential threat: Americans who may not want the consequences in their backyards.
And growing opposition could turn into a bigger financial problem.
Read more: Palo Alto CEO wants AI labs out of the ‘fear-mongering business’
Americans are souring on the data centers powering the AI boom
Wall Street has spent the bulk of the past three years worrying if the AI boom has gotten too expensive. Dario Perkins thinks investors might be missing a different kind of problem.
“AI boosters on social media live in a bubble,” Perkins wrote in a Monday research note as reported by Business Insider. His blunt remark serves as a reminder about Americans unhappy with the industry’s expansion: “And they vote.”
That could sound like an unusually political risk for a Wall Street forecast, but the public-opinion numbers are tough to dismiss.
In a Pew Research Center survey released in September, 60% of Americans said they’d be uncomfortable with a new data center operating in their area. Just 7% said they’d be very or extremely comfortable.
More Artificial Intelligence:
- AI’s biggest business opportunity may not look like AI
- AI racks are outgrowing old power: 2 chipmakers have a strategy
- BofA splits software into winners and losers as AI agents arrive
However, the sentiment has soured even more, as in January, 30% thought data centers were mostly bad for the quality of life of nearby residents. By August, that share reached 49%. The percentage who said data centers were bad for home energy costs rose from 38% to 50%.
And the concern goes beyond electricity bills.
In a separate Pew survey, 71% of Americans expected AI to reduce U.S. jobs over the next two decades, up from 64% in 2024.
That said, polling alone doesn’t stop construction. But in some communities, objections have already turned into delays and cancellations.
Local opposition threatens a massive bet on AI infrastructure
The backlash is showing up in development plans already.
Goldman Sachs says 86 U.S. data-center projects have been canceled or extensively delayed since January, with 76 linked directly to local opposition. So clearly, things have gotten past the point of angry comments at town meetings. It’s a potential problem for companies counting on those facilities to open on schedule.
And it’s easy to speculate on why neighbors are paying attention.
Consider a 100-megawatt data center running at full capacity around the clock. It would consume 876 million kilowatt-hours a year, about as much electricity as 83,000 average American households, based on federal consumption figures.
Still, power costs aren’t hypothetical. PJM, the grid operator serving parts of 13 states and Washington, D.C., said through its independent market monitor that existing and expected data-center demand accounted for 9% of wholesale electricity costs through July 2026.
For Perkins, this is exactly where a neighborhood fight becomes Wall Street’s problem.
Building a data center entails years of spending before a customer starts paying for its computing power.
Delay an opening, and the financing costs don’t disappear. Perkins estimates debt across parts of the AI ecosystem could approach $2 trillion by the end of 2027. S&P Global has separately warned that complex AI financing deals might leave some obligations less visible to investors.
There is a catch to the backlash narrative, though. SemiAnalysis found that many local moratoriums have barely affected projects already underway; roughly 2.3 gigawatts of the 21 gigawatts in affected areas are likely to face delays.
AI construction isn’t stopping. But companies that can secure power, permits, and affordable financing may be the ones best positioned to weather the resistance.
Electricity costs and borrowing risks could decide AI’s next chapter
Perkins comments leave Big Tech with some difficult choices.
Companies could shoulder more of the costs of power infrastructure, while utilities introduce special rates for large data-center customers. Developers might also seek locations with fewer electricity constraints or secure dedicated power supplies.
That said, none of these solutions comes for free, and many of the Big Tech giants are under immense pressure already.
According to FactSet, Microsoft (MSFT), Alphabet (GOOGL), Amazon (AMZN), Meta Platforms (META), and Oracle (ORCL) are expected to drop more than $690 billion on capital expenditures in fiscal 2026, relying a ton on external financing as their AI investments strain free cash flow.
For investors, the stakes extend well beyond Nvidia’s next chip orders. Rising borrowing costs, mounting project delays, and changes to spending plans at Microsoft, Amazon, Alphabet, and Meta could signal early trouble.
Equally important is whether cloud-related sales and demand for paid AI services grow quickly enough to justify the enormous investments already underway.
A slowdown wouldn’t necessarily mean AI has failed. It could simply mean the financial and physical costs of expansion are becoming harder to ignore.
Wall Street is betting that AI will transform the American economy.
Perkins’ warning suggests ordinary Americans may have more say in how quickly that future arrives than investors realize.
Related: Bank of America sees a costly catch in America’s strong economy
