The rapid growth of artificial intelligence is creating a new conflict in the United States: tech companies are building more and more data centers, promising jobs and economic growth, while city residents fear that along with them will come additional strain on the power grid, water supply, the environment, and their own health.

This conflict is especially acute in areas that have already spent decades living next to major industrial facilities. For their residents, the construction of yet another site tied to massive energy use and infrastructure looks not like a symbol of technological progress, but like a return to a painfully familiar past, TechCrunch reports.

“We Already Know How This Ends”

In Philadelphia, one of the leading figures in the movement against new data centers is Shomar Pitts, who has lived his entire life in the Grays Ferry neighborhood in the city’s southwest.

His childhood was spent next to one of the largest oil refineries on the East Coast — Philadelphia Energy Solutions. The refinery processed up to 335,000 barrels of oil per day and shut down after a major explosion in 2019.

Pitts recalls that for a long time he did not connect the chronic illnesses of his relatives and neighbors with industrial pollution. In 2018, he met activists from the environmental organization Philly Thrive, who were telling residents about the consequences of living near an oil refinery.

“My mother has this, my sister has that, my next-door neighbor suffers from this,” Pitts told TechCrunch. After that, he joined the movement and today serves as managing co-director of Philly Thrive.

The organization is now calling for a moratorium on the construction of data centers in Philadelphia.

There are currently no official construction projects in the city, but authorities have already identified two potential sites. One is in Northeast Philadelphia, and the other is in Grays Ferry — the very neighborhood that spent decades living next to an oil refinery.

For local activists, the issue is not only the scale of pollution. They do not want the neighborhood to once again become a site for major infrastructure whose consequences residents will feel for years.

A Data Center Is Not an Oil Refinery, but It Still Needs a Lot of Energy

It would be inaccurate to directly compare a modern data center with an oil refinery. The Philadelphia refinery was a much larger industrial facility with an obvious environmental impact.

But the artificial intelligence industry is rapidly increasing its energy consumption, and that is exactly what is raising more and more questions.

According to a BloombergNEF forecast, by 2035 U.S. data centers could consume more natural gas than Germany and Japan combined. Moreover, this new estimate is nearly twice as high as the forecast the company made just nine months ago.

Natural gas emits fewer pollutants when burned than oil, but it is still a fossil fuel and is still associated with carbon dioxide emissions.

In addition, data centers need backup power sources. As environmental activists note, a typical data center project in Pennsylvania may include hundreds of diesel generators. Formally, they are intended for emergencies, but during periods of high strain on the power system, they are already being proposed for use to reduce pressure on the grid.

This creates an additional problem: the more data centers appear, the greater the strain on infrastructure may become, and maintaining that infrastructure may also rely in part on fossil fuels.

According to Annie Fox, an attorney at the Clean Air Council, rising demand from data centers could intensify climate change and, in turn, contribute to more frequent and more severe heat waves, when the power system is already under maximum strain.

The Main Question Is Who Will Pay for the AI Boom

For residents of American cities, the debate over data centers is gradually turning from a question about technology into a question about how resources are allocated.

Data centers require electricity, water, and land. At the same time, companies promise new jobs and economic activity, but residents do not always see these benefits as sufficient compensation for the potential strain on infrastructure.

At a rally in Philadelphia, local resident Donna Greenberg spoke not only about pollution, but also about water, electricity, and privacy. According to her, she is concerned that new facilities may use resources that local neighborhoods need.

This problem extends far beyond Philadelphia. Protests against the construction of data centers are taking place in different states, especially in places where local resources are limited or residents already have experience living next to major industry.

In some cities, authorities have already begun introducing temporary restrictions. Moratoriums on data center construction have been adopted, among other places, in Denver, Indianapolis, Asheville, Charlotte, and Reno. Usually, the goal of such restrictions is not to ban the technology entirely, but to give authorities time to study the potential consequences of construction and determine the necessary rules.

In July, New York Governor Kathy Hochul also signed an order temporarily prohibiting the state from approving permits for major new data center projects.

“Progress shouldn’t arrive with a higher utility bill, deleted water supply, or noise pollution,” Hochul said. According to her, data centers should be built where people actually want to see them.

Why the Conflict Has Become So Acute Right Now

For residents like Pitts, the issue is not just about a specific building or the number of servers. It is a matter of trust.

Neighborhoods that have already endured decades of living next to heavy industry know that promises of economic benefit do not always mean the absence of long-term consequences. That is why the argument that AI data centers will create jobs and bring investment may sound far less convincing to such communities than it does to areas without that historical experience.

In Philadelphia, more than one hundred people came to the first rally of the No Data Centers in Philly campaign. Activists demanded that any possible placement of such facilities in the city be stopped.

At the same time, this standoff does not necessarily mean rejecting the development of AI infrastructure altogether. The main question for local authorities and residents is where exactly it should be built, what resources it will consume, what environmental consequences are acceptable, and who should bear the associated costs.

For now, the AI industry is expanding infrastructure at a pace that is hard to compare with previous technological cycles. And now it has to confront a limitation that cannot be solved solely with more powerful chips: local communities may simply not want new computing capacity near their homes.

For Philadelphia, this story is especially symbolic. A neighborhood that only recently lived next to a massive oil refinery is now discussing the possibility of living alongside the infrastructure of another era — no longer oil, but AI.