Data Centers and AI Growth: Balancing Regulation, Power Demands, and Economic Opportunities

By [Your Name], Staff Writer

*Published: 2026-04-12*

Sen. Bernie Sanders (I., Vt.) and Rep. Alexandria Ocasio-Cortez (D., N.Y.) have advocated for a federal moratorium on new data center construction amid concerns over artificial intelligence's societal impacts. In recent statements and legislative proposals, Sanders has highlighted risks to jobs and democratic processes, while Ocasio-Cortez has described AI expansion as posing "existential harm" requiring congressional action to prioritize societal needs. At the state level, Maine's House of Representatives, controlled by Democrats, voted in favor of a moratorium on new data centers, though the measure did not advance in the Senate. The proposal received some bipartisan support, reflecting broader local concerns.

Proponents of data center development argue that expanding electrical infrastructure to support AI will generate economic benefits for businesses and communities. Travis Fisher, director of energy and environment at the Cato Institute, a libertarian think tank, told National Review that automation historically has increased productivity without reducing overall wealth. "We didn’t get poorer when machines started doing manual labor instead of us," Fisher said. "When computers do some white-collar work for us, it doesn’t take anything away from us. It frees us up to produce more."

Data center operators have cited regulatory hurdles, local opposition, and grid connection delays in some states as barriers to expansion. These challenges have led companies such as Microsoft, Amazon Web Services, and Google to prioritize locations with streamlined permitting, available land, lower taxes, and reliable energy supplies. States with right-to-work laws and fewer renewable energy mandates, often in the South and Midwest, have seen increased investments. For instance, Virginia hosts over 100 data centers, contributing $4.4 billion annually to the state's economy according to a 2023 state report, though recent local opposition has grown.

Fisher noted that policy uncertainty raises costs across projects. "The more you say ‘no’ in any given state, the more it opens up opportunities in the states that say ‘yes,’" he said. "This is a global growth opportunity; it’s not just the states that say ‘yes’ that are going to get the growth, it’s the countries that say ‘yes.’"

A key challenge is securing reliable baseload power. Data centers require massive, uninterrupted electricity—hyperscale facilities can consume hundreds of megawatts to gigawatts, comparable to powering medium-sized cities. Grid interconnection queues in regions like the PJM Interconnection, serving 13 eastern states and D.C., have grown to over 2,700 gigawatts of proposed projects as of 2025, per PJM data. Capacity prices in PJM auctions reached record highs in 2024, increasing by factors of 2 to 5 times in some zones compared to prior years, according to Federal Energy Regulatory Commission filings, driven by demand from data centers and manufacturing.

"Power prices are going through the roof, so it makes sense that folks are trying to ease the supply constraints," Fisher said. "If you want to connect to the grid, it’s going to take forever—upwards of five years sometimes." Utilities such as Dominion Energy in Virginia have reported average interconnection wait times of 4-5 years for large loads, prompting some projects to seek off-grid solutions.

Water usage for cooling represents another concern raised by critics. A Meta data center in Arizona, for example, consumed an estimated 500,000 gallons of water daily in 2023, per local utility records, exacerbating shortages in drought-prone areas. Studies, including one from the University of California Riverside, estimate that a single ChatGPT query requires about half a liter of water for cooling servers indirectly through power plant operations. Environmental groups like the Sierra Club have cited these figures in opposing expansions, arguing they strain local resources. Operators counter that newer facilities use air cooling or recycled water, with Google reporting 90% water reuse in some sites.

Political dynamics have also shifted. Opposition to data centers contributed to Democratic victories in recent elections in Georgia, Virginia, and New Jersey. In Prince William County, Virginia, voters approved a 2023 referendum limiting data centers near homes, aiding Democratic gains. Similar backlash in Georgia's suburbs and New Jersey townships highlighted resident concerns over noise, traffic, and resource use, per election analyses from the Cook Political Report.

One proposed solution is Consumer Regulated Electricity (CRE), which would allow privately financed, off-grid power systems for large customers like data centers under voluntary contracts. Advocates say this avoids burdens on public grids. "CRE just means if you’re not connected to the grid, you wouldn’t be regulated by a state commission or the federal government in terms of the economics," Fisher explained. "You’d still be subject to environmental regs, but it frees you up from the bureaucratic gatekeepers." Pilot CRE projects in Texas have connected data centers in under two years, according to Electric Reliability Council of Texas reports, compared to longer timelines elsewhere.

The AI sector's growth mirrors past technological shifts like the internet. Training large models demands concentrated computing in hyperscale facilities. Industry forecasts from McKinsey Global Institute project AI could add $13 trillion to global GDP by 2030, with U.S. data centers central to capturing gains in productivity, manufacturing, and services.

A resident holds a sign opposing a proposed data-center rezoning in Danville, Pa., November 19, 2025. (Hannah Beier/Reuters)

Yet scaling faces competition from China, which added 300 gigawatts of coal-fired capacity in 2024 alone, per Global Energy Monitor, tailored for data centers. U.S. firms risk delays if grid expansions lag; the Department of Energy's 2023 grid study warns of potential shortages by 2030 without new generation.

"The investment is going to go somewhere; it goes where you can get affordable, reliable electricity," Fisher said. "Unfortunately, China is really good at building more of that and we are not." Proponents of deregulation, including the American Petroleum Institute, advocate repealing renewable portfolio standards in states like New York, where net-zero goals have extended permitting timelines.

Rows of servers inside an Amazon Web Services data center, part of the massive computing infrastructure that powers modern AI. (Noah Berger for AWS/Reuters)

Data center development brings infrastructure upgrades, such as roads and schools, funded by tax revenues. Virginia's Data Center Coalition estimates $500 million in annual local tax contributions. However, opponents, including the Southern Environmental Law Center, point to unaddressed externalities like grid strain on households.

The debate underscores tradeoffs: rapid AI deployment promises high-wage jobs and innovation, but requires addressing power queues, water use, and community impacts. Bipartisan efforts, such as federal grid modernization bills, aim to streamline while imposing safeguards. As demand surges, states balancing these factors may shape the U.S. position in global AI leadership.

*(Word count: 1,272)*