Growing resistance to data center projects
AI data centers are spreading across the United States at a remarkable pace, but there’s a problem brewing. Local communities are starting to push back, and this resistance could actually slow down the infrastructure growth that artificial intelligence depends on. I think we’re seeing something important here—people are asking questions about what these massive facilities mean for their towns and cities.
A new report from the Brookings Institution points to specific concerns that keep coming up. Electricity use is a big one, obviously. Then there’s water consumption, which many people don’t realize can be substantial for cooling these centers. Tax abatements and environmental impact round out the main complaints. The authors argue that without addressing these issues properly, local opposition might constrain how quickly AI infrastructure can expand.
Why community voices matter
Shaolei Ren, an electrical engineering professor at UC Riverside, makes a point that seems obvious once you hear it. Community satisfaction should be the metric that really matters. It’s not just about technical specs or economic benefits—people need to feel good about what’s being built in their backyard. Ren told Decrypt that there’s growing recognition that local voices should shape these projects from the beginning.
“You cannot improve what you do not measure,” he added, which strikes me as both simple and profound. If companies don’t track how communities feel about data centers, how can they make them better neighbors?
The scale of the challenge
Let’s talk numbers for a moment. U.S. data centers consumed about 183 terawatt-hours of electricity in 2024. That’s roughly equivalent to Pakistan’s entire annual energy demand. When you put it that way, local concerns about electricity costs and availability start to make more sense.
The Brookings report mentions something called a “techlash” against the AI sector. People worry about jobs disappearing due to automation, plus all that energy consumption and environmental impact. These anxieties have sparked protests in various communities. Left unaddressed, the report warns, these concerns could slow data center construction, weaken AI growth, and limit the benefits that tech firms and government officials keep promising.
Where development is happening
Much of the new data center development is concentrated in the American South. Companies are building large facilities in North Carolina, Georgia, Virginia, Louisiana, Mississippi, Alabama, South Carolina, and Tennessee. But here’s the thing—local leaders and advocates argue that these centers often end up in low and middle-income areas. Places that might lack the political influence to say no or negotiate better terms.
Major tech firms aren’t slowing down, though. Amazon and Nvidia have announced multibillion-dollar investments to expand data center and AI infrastructure. According to Data Center Map, there are nearly 4,000 data centers in the U.S. and about 10,700 worldwide. That network keeps growing.
A possible solution
Brookings suggests legally binding community benefit agreements as an alternative to informal negotiations. These CBAs would define costs, subsidies, and tax revenues upfront. They’d also set enforceable commitments for jobs, electricity and water use, and pollution controls.
The thinking is that greater transparency on these fronts would help ease public worries. Well-crafted agreements could address known problems before construction even begins. Perhaps this approach represents a middle ground—acknowledging that data centers are necessary for AI development while ensuring communities get fair treatment.
It’s worth noting that in January 2025, the Trump administration announced Stargate, a $500 billion AI infrastructure initiative backed by OpenAI and Oracle. That announcement included calls for long-term safeguards to ensure communities aren’t left with data centers that deliver little benefit. So the conversation is happening at multiple levels.
The basic tension seems clear. Data centers are controversial but critical. Without enough of them, the digital revolution could potentially stall. But building them without community support creates its own problems. Finding the right balance might be one of the bigger challenges for AI infrastructure in the coming years.
