AI companies need massive data centers to train their models. Local communities across America don't want them. And someone's trillion-dollar bet is about to hit a wall made of zoning boards and concerned citizens.
Fortune reports that communities nationwide are blocking billions in data center projects, even as Big Tech has collectively wagered over $1 trillion on building them. Multiple cities have now voted to ban or severely restrict data centers, creating a showdown between infrastructure needs and local opposition.
The complaints are consistent: data centers consume enormous amounts of electricity, strain local power grids, require massive amounts of water for cooling, generate noise from cooling systems, and provide relatively few local jobs once operational. They're industrial facilities dressed up as tech infrastructure.
In Northern Virginia, the global capital of data centers, communities are pushing back against further expansion. In the Midwest, towns that initially welcomed data centers as economic development are now having second thoughts. And in places like Idaho and New Jersey, proposed facilities have been outright rejected.
Here's the fundamental tension: AI training requires compute at scales that are genuinely unprecedented. A single frontier model training run can consume more electricity than a small city uses in a year. That compute has to live somewhere. And "somewhere" increasingly means purpose-built data centers in places with cheap power and favorable tax treatment.
But those places are discovering that data centers aren't the economic windfall they were promised. Once built, they employ dozens of people, not thousands. They strain infrastructure without generating proportional tax revenue. And they create environmental impacts that residents don't want.
The irony is that many of these facilities are being marketed as "green" because they'll eventually use renewable energy. But in practice, they're driving up power demand so fast that utilities are keeping coal plants online longer and building new natural gas capacity. The renewable energy promises are aspirational. The power consumption is immediate.
