Alan Armstrong

Oklahoma Republican Senator Alan Armstrong is pushing Congress to overhaul federal permitting rules, arguing in a Newsweek interview that the country needs new power infrastructure to accommodate data centers and other growing electricity demand without passing costs on to households.

His argument comes as artificial intelligence infrastructure becomes an increasingly potent affordability issue: 71 percent of Americans oppose building an AI data center in their local area, according to Gallup, while researchers debate whether the rapid buildout is already driving up household electricity prices.

Armstrong, appointed to the Senate this year and unable to run for a full term under Oklahoma law, told Newsweek that data centers have become a convenient target for frustration over utility bills. He instead blamed inadequate generation and grid infrastructure and said his permitting proposal would require large new users to shoulder the additional costs they impose on the system.

“I think that we have to have data centers,” Armstrong said in the interview with Newsweek. “I think data centers have kind of become the scapegoat on high energy prices.”

The issue is especially significant in Armstrong’s home state, where proposed data centers have encountered objections over electricity costs, water consumption, noise, and changes to surrounding communities.

Public resistance to data centers is substantial and crosses party lines.

Gallup's March 2-18 survey found that 71 percent of U.S. adults opposed building an AI data center in their local area, including 48 percent who strongly opposed it. Only 7 percent strongly favored it. Opposition was higher than the 53 percent who opposed building a nuclear power plant in their area in the same survey. The poll of 1,000 adults had a sampling error margin of plus or minus 4 percentage points.

The reasons extend far beyond generalized discomfort with AI. In a separate April Gallup survey, half of data-center opponents raised concerns about resource use, including 18 percent who specifically cited water and another 18 percent who mentioned energy. Roughly one in five opponents pointed to potential economic consequences, including higher utility bills and cost-of-living increases. Gallup found majorities across major political groups opposed local construction.

Armstrong rejected the idea that the solution is restricting the industry's growth.

“I don't think anybody's ever been successful in civilization at stopping progress,” he told Newsweek. “And I don't think that's a smart thing for us to do.”

He said the government should set clear rules and safety standards for AI while keeping an environment attractive to companies developing the technology, arguing that domestic attempts to stop the industry risk pushing investment elsewhere.

Armstrong's contention that data centers are being unfairly blamed for higher electricity bills touches on a rapidly developing economic debate.

Few dispute the scale of the new demand. An August 2026 analysis by Lawrence Berkeley National Laboratory and the Brattle Group said U.S. data-center electricity demand more than doubled from 2018 to 2024 and could roughly triple again from 2024 through 2028. The researchers warned that serving very large customers creates financial and operational risks for utilities, including insufficient power supply and investments that could go unused, with consequences that could reach other customers.

Research released this year, however, reaches strikingly different conclusions about how that growth affects household bills.

A September analysis by Cameron Scalera, Valentina Bosetti, and Filippo Pecci examined roughly 1,200 utility territories and about 2,700 operating data centers. The researchers found that additional data centers increased residential electricity prices, with households experiencing larger increases than commercial or industrial customers. Their estimate attributed a roughly 4 percent residential price increase to an average-sized hyperscale facility. However, the effect was close to zero in territories served by nonprofit utilities.

That study points to an important distinction in the affordability argument: the question is not simply how much electricity a data center consumes, but whether sufficient capacity already exists and how utilities assign the cost of new infrastructure. A system with spare capacity may absorb additional demand relatively cheaply. Where capacity is constrained, large new loads can require more expensive generation, transmission or distribution improvements whose costs may ultimately reach other customers.

Scalera, a Columbia University researcher whose team has studied data centers and electricity prices, told Newsweek that emerging evidence suggests the facilities are pushing residential rates higher.

“We find that data centers do cause household electricity prices to rise,” Scalera said. “We find that about 10 data centers lead to a 11-13% increase in prices on average, and a single large hyperscaler causes prices to rise by about 4% on average.”

Scalera said Armstrong’s argument about inadequate infrastructure has some merit because his team found price increases were less severe in areas with greater surplus capacity. But he added that data centers themselves remain “an independent contributor to rising electricity prices” and said a key question is who ultimately pays for necessary grid upgrades.

On Armstrong’s proposal to use large-load tariffs to protect residential customers, Scalera said the policies are still too new to judge conclusively. “Maybe. But we do not know for sure,” he said. “They seem reasonable. Future research will tell whether they actually help to reduce household rate increases.”

Other researchers have reached almost the opposite conclusion. A 2026 Electric Power Research Institute study found that from 2015 through 2024, data-center growth modestly lowered average retail electricity rates rather than increasing them. Its researchers estimated that doubling data-center capacity reduced residential electricity prices by about 3.5 percent, arguing that utilities can spread fixed system costs across more electricity sales.

Importantly, EPRI cautioned against simply applying that historical finding to the coming AI expansion. The study acknowledged that supply constraints, the unprecedented size of hyperscale facilities, and investments made for demand that later fails to materialize could change the outcome.

Armstrong told Newsweek that under his proposal, data centers would bear both the incremental costs their demand creates and existing charges.

“The data centers will be obligated to make sure they pay not only their fair share, but actually a little bit more than their fair share for power,” Armstrong said, arguing the arrangement could help lower other customers' utility bills.

Large-load tariffs designed to isolate those expenses are already becoming an important regulatory tool. Berkeley Lab's review of 55 large-load tariffs found provisions including minimum monthly bills, demand charges, collateral requirements and direct assignment of costs. The last mechanism is explicitly intended to prevent existing customers from subsidizing infrastructure required by new large loads.

Armstrong was more openly critical of President Donald Trump when asked about Russian diesel, arguing the United States should expand domestic infrastructure instead of directing money toward Moscow.

“I don't like the idea of us sending revenues to, I think, one of our very worst adversaries and particularly don't like the message we're sending to the Ukrainian people who, you know, have relied on us,” Armstrong told Newsweek. “And so I don't like that message. And I think there are much better ways for us to get diesel prices lower here in the US.”

Armstrong tied the question to his broader permitting agenda, arguing that restrictions and delays in pipelines and refining capacity have limited America's abil