Americans are more worried about AI data centers moving into their neighborhoods than they are about nuclear power plants, according to recent polling, and the gap reveals a deeper problem: developers aren't being transparent about what these massive facilities will actually demand from local communities. A Gallup survey found that 71% of Americans oppose data center construction locally, compared to 45% who would accept a nuclear plant. The opposition is driven less by fear of the unknown and more by specific, measurable concerns that communities say they can't get clear answers about.

What Are Hyperscale Data Centers, and Why Are Communities So Concerned?

Hyperscale data centers are massive computing facilities that meet three criteria: they use at least 100 megawatts of electricity (enough to power roughly 80,000 U.S. households), contain at least 5,000 computer servers, and occupy at least 10,000 square feet. These aren't small operations. An AI prompt can require 10 to 15 times as much energy as a Google search, and training large AI models requires 10 to 40 times more power than traditional data centers, according to research from the University of Virginia.

The scale of expansion is staggering. The U.S. currently has approximately 5,400 data centers, and that number could double by 2030. Close to 3,000 new data centers have been proposed nationwide. Global power demand for data centers is expected to nearly double from 103 gigawatts to 200 gigawatts by 2030, requiring an estimated $3 trillion in investments.

Communities across the country are mobilizing against these projects. In Dubuque County, Iowa, approximately 300 people attended a recent town hall on data centers, a turnout that dwarfs typical civic engagement on other issues. This groundswell of opposition is occurring town-to-town and county-to-county across the nation.

Why Is Transparency the Real Issue Driving Opposition?

While energy and water consumption dominate public concern, the South Dakota Legislative Research Council identified a critical barrier to informed decision-making: non-disclosure agreements between developers and local governments prevent the public from accessing information about projected resource demands, impacts, and scale of data center projects. A 2025 Harvard Law School study found that this lack of transparency in power rate contract proceedings results in data centers paying discounted rates, which ultimately forces residential utility rates up for everyone else.

The LRC concluded that "the best possible outcomes can be achieved only if there is an insistence on transparency regarding the impacts of data centers". South Dakota has already taken one step forward: Senate Bill 135, signed into law in March 2026, requires data center operators to report projected water usage to local water providers and the Board of Water Management before operations begin, followed by public, semi-annual water usage reports. Utah and Virginia approved similar water-reporting requirements this year.

Without transparency, communities are left guessing about real impacts. One data center using evaporative cooling for server cooling would have a water usage rate equivalent to the entire demand of either Pierre or Sioux Falls, South Dakota. Google's data center in Council Bluffs, Iowa, uses water equivalent to the demand of four million households.

What Specific Concerns Drive Opposition to Data Centers?

Gallup's breakdown of opposition reasons reveals that battery storage and pollution-reduction measures address only a fraction of community concerns. Here's what opponents actually worry about:

• Water and Energy Consumption: Roughly 36% of opponents cite excessive resource use, split nearly evenly between water consumption at 18% and energy consumption at 18%. These are the largest single categories of concern.
• Pollution and Noise: Only 16% of opposition traces to pollution, air quality, and noise concerns, the category that battery energy storage systems most directly address.
• Quality of Life and Economic Concerns: About 20% cite quality-of-life issues such as traffic, population growth, and competing land use preferences, while a comparable share raise economic concerns, including fears that data centers will drive up household electricity costs.

The distinction matters because battery storage systems, while valuable for managing grid interconnection and reducing diesel generator noise, do not reduce a facility's total water draw for cooling or total electricity consumption. A lithium iron phosphate battery system built to manage AI-driven load fluctuations smooths how a facility draws power from the grid, but it doesn't change the aggregate load a hyperscale campus places on a regional grid.

Noise pollution is a genuine concern in communities near data centers. Thousands of servers operating together can raise noise levels to 96 decibels. Backup diesel generators can reach up to 105 decibels, as loud as a jet flying overhead. At an xAI facility in Southaven, Mississippi, 27 natural gas-powered turbines operating 24/7 are disturbing the sleep of nearby neighbors. Cooling towers can emit up to 70 decibels within 400 feet.

How Can Communities and Policymakers Address Data Center Concerns?

Experts and policymakers are converging on several practical steps to balance AI infrastructure development with community protection:

• Require Public Disclosure of Resource Demands: Mandate that developers publicly report projected water usage, electricity consumption, and environmental impacts before projects receive approval. This allows communities to make informed decisions and utilities to plan infrastructure upgrades without stranding assets if data centers cease operations.
• Establish Financial Safeguards for Stranded Assets: Infrastructure for electricity and water can last 30 years. If data centers reduce or cease operations before upgrade costs are recovered, residential ratepayers could end up paying for unnecessary infrastructure. Texas passed a law in 2025 setting standards for stranded assets and financial commitments from data centers.
• Diversify Cooling Technologies and Power Sources: Data centers currently get 40% of their electricity from natural gas, 24% from renewables, 20% from nuclear, and 15% from coal. Emerging technologies like liquid immersion cooling could use less water and less electricity than evaporative cooling, though they require specialized equipment.
• Implement Air and Noise Quality Standards: Natural gas turbines and diesel generators used by data centers can exacerbate air quality problems and create noise pollution. Communities should establish clear ordinances and permitting requirements for on-site generation.

Some communities are already seeing tangible benefits. A parish in Louisiana is giving teachers $50,000 bonuses via revenue from a planned $27 billion Meta data center, demonstrating that deep-pocketed developers can provide meaningful community benefits.

The tension between AI infrastructure needs and community concerns is unlikely to resolve quickly. President Donald Trump and Senate Intelligence Committee Chairman Tom Cotton have emphasized that maintaining America's AI advantage is vital to economic strength, national security, and military power. However, Cotton also acknowledged that "Americans certainly have valid concerns about potential rising energy costs and strains on natural resources related to data centers". The challenge ahead is ensuring that transparency, not secrecy, guides how communities and developers navigate this critical infrastructure buildout.