America's AI boom powering electricity squeeze

When Kurt Borchardt opened his latest electricity bill, he thought there must have been a mistake.

"Our electric bill doubled in one month. Almost a $3,000-$4,000 jump on a single bill," wrote Borchardt, co-owner of Artisanal Brew Works in Saratoga Springs, New York, describing the shock on social media.

The brewery had already endured a slow winter season, traditionally its weakest period. Then came what he said was a 133-percent increase in the electricity supply rate. The company's National Grid bill has now become its second-largest expense after rent, squeezing margins at a time when customer traffic remains slow.

"When I saw that bill, I fell out of my chair," Borchardt told local television station WTEN. His frustration reflects a broader national trend.

Electricity prices in the United States are emerging as a new source of economic strain, raising concerns about inflation, industrial competitiveness and political risk, particularly after a colder-than-average winter increased heating demand and tightened natural gas markets.

The most recent US Bureau of Labor Statistics Consumer Price Index report showed that overall inflation rose 2.4 percent in the 12 months ending in January, while electricity prices increased 6.3 percent. Though gasoline prices have fluctuated, electricity bills have continued to climb steadily, placing sustained pressure on both households and businesses.

One key factor behind the rise in electricity prices is surging power demand from data centers and artificial intelligence applications. As the US accelerates investment in AI infrastructure, electricity consumption from large-scale computing facilities has expanded rapidly, placing additional strain on an already aging power grid.

"Since electricity is a very inelastic good, these price increases will continue to put upward pressure on inflation," Aaron Pacitti, a professor of economics at Siena University, told China Daily. "One of the main drivers of this increase is the rise in electricity demand from data centers and increased usage of AI."

According to Lawrence Berkeley National Laboratory, data centers accounted for about 4.4 percent of total US electricity consumption in 2023. Depending on the pace of broader economic growth, that share is projected to rise to between 6.7 percent and 12 percent by 2028.

In absolute terms, total data center electricity use climbed from 58 terawatt-hours (58 trillion watt-hours) in 2014 to 176 TWh in 2023 and could reach between 325 and 580 TWh by 2028.

For comparison, the average US household consumes roughly 10 to 11 megawatt-hours (MWh) of electricity per year, according to the US Energy Information Administration. That means 176 TWh is equivalent to the annual electricity consumption of roughly 16 million households, while 580 TWh would equal the yearly power use of more than 50 million homes.

In other words, by 2028, electricity demand from data centers alone could rival the total residential electricity consumption of dozens of US states.

Similar challenges are emerging in other major technology markets as governments seek to balance the rapid growth of artificial intelligence with the need for reliable power supply.

In China, electricity demand from data centers is also expected to rise rapidly as artificial intelligence development accelerates. Authorities have encouraged new data centers to be built in regions with abundant renewable energy resources and lower electricity costs as part of broader efforts to manage growing computing demand.

"China already generates more than twice as much electricity as the United States and has increased its total power generation by nearly 6 percent per year over the past decade," said Kyle Chan, a researcher focusing on global energy and technology policy at the Brookings Institution. "Over half of China's electricity growth during this period has come from clean energy sources such as wind, solar and hydropower."

In the United States, the surge in electricity demand is already beginning to show up in capacity markets.

PJM's latest capacity auction for the 2027/2028 delivery year fell 6,623 megawatts short of its reliability requirement, underscoring a growing imbalance between electricity supply and demand, according to a Dec 17 news release from the grid operator, which serves 13 states and the District of Columbia.

Capacity auctions are forward-looking markets in which grid operators secure commitments from power plants to ensure sufficient supply during future peak demand periods.

"But this auction leaves no doubt that data centers' demand for electricity continues to far outstrip new supply, and the solution will require concerted action involving PJM, its stakeholders, state and federal partners, and the data center industry itself," said Stu Bresler, executive vice-president of market services and strategy at PJM.

Economists warn that persistently higher utility costs could weigh on overall economic momentum.

For manufacturers, especially energy-intensive sectors, higher electricity prices translate directly into rising production costs.

"Higher energy costs will act as a drag on growth and competitiveness for US firms and heighten the affordability issues facing US households," Pacitti said. "Since demand from data centers and AI is unlikely to subside anytime soon, these price increases will act as a modest headwind to growth."

Beyond demand growth, structural challenges are also contributing to the problem. In many parts of the country, utilities purchase electricity through wholesale markets, and when demand rises faster than supply, prices increase for all consumers, according to Ari Peskoe, director of the Electricity Law Initiative at Harvard Law School.

In addition, when utilities build new infrastructure — such as transmission lines or power plants — the costs are typically spread across all ratepayers. "When a utility builds infrastructure like poles and wires or power plants, it socializes that cost across all users," Peskoe told Harvard Law Today, meaning consumers ultimately share the cost of expanding the system to accommodate large-scale computing facilities.

The Department of Energy has repeatedly pointed out that more than 70 percent of transmission lines are over 25 years old and require substantial upgrades. Investment has not always kept pace with evolving needs, and extreme weather events have added further strain.

"It is not only a bottleneck for growth, but also a threat to economic and national security," Pacitti said. "Firms and households rely on a steady stream of access to power at low cost. An aging infrastructure, coupled with rising demand from technological applications, puts that at risk."

Some energy policy experts argue that large data center operators could ease pressure on the public grid by financing or developing dedicated power generation for their facilities. However, such proposals face political and regulatory hurdles.

According to projections from Lawrence Berkeley National Laboratory and other energy analysts, electricity demand from data centers is expected to remain elevated in the coming years. Without significant acceleration in new generating capacity and grid investment, electricity prices are likely to remain under pressure, influencing both economic conditions and political debates in the United States.

bilinlin@chinadailyusa.com