The rapid expansion of power-hungry artificial intelligence data centers has already increased wholesale electricity prices by an estimated 2% to 6% nationwide, with researchers warning that generation costs could rise far more sharply by 2028 as new facilities connect to the grid.
By yourNEWS Media Newsroom
The artificial intelligence boom is beginning to show up in the cost of electricity, with new research indicating that the rapid construction of data centers has already placed upward pressure on wholesale power prices and could contribute to higher household utility bills in the years ahead.
Researchers at the Federal Reserve Bank of Dallas estimate that existing data centers have raised average wholesale electricity prices by approximately 2% to 6% nationwide, with significantly larger increases in regions where the facilities are heavily concentrated.
The pressure could intensify as technology companies build more facilities to support increasingly powerful artificial intelligence systems.
Under the researchers’ preferred middle-range projection, the cost of generating electricity could be approximately 20% to 30% higher by 2028 than it would have been without the additional demand created by new data centers.
That projection does not mean residential electricity bills themselves are expected to rise by 20% to 30%.
Wholesale power costs represent only one component of a household electric bill. Consumers also pay for transmission, distribution and other utility expenses, and changes in wholesale electricity prices can take time to reach retail customers.
Dallas Fed researchers estimate that energy costs account for roughly half of a typical retail electricity price, making the ultimate effect on individual households dependent on a broader set of factors.
Even so, the research suggests electricity costs are likely to face sustained upward pressure over the next two years as utilities adapt to extraordinary new levels of demand.
The underlying challenge is the enormous quantity of power required to operate modern data centers.
Depending on its size and computing capacity, a single large data center can consume approximately as much electricity as a small city, according to estimates compiled by Dallas Fed researchers.
wp2606 by yourNEWS Media
Artificial intelligence workloads require vast numbers of processors operating continuously, while data centers also need substantial additional electricity for cooling systems and supporting infrastructure.
As more facilities connect to existing power networks, utilities can be forced to expand generating capacity and strengthen the electrical infrastructure needed to deliver that power.
Those investments can include new power plants, transmission lines, substations and other equipment required to accommodate the sudden arrival of exceptionally large electricity users.
A central question is who ultimately pays for those upgrades.
State regulators and utilities must decide how infrastructure costs are divided between the companies operating data centers and existing residential and commercial customers. Those decisions could substantially affect how much of the AI industry’s growing electricity demand eventually appears on household bills.
The issue has increasingly moved from utility planning meetings into state and national politics as elected officials attempt to promote artificial intelligence development without imposing higher energy costs on residents.
President Donald Trump has supported expanding U.S. artificial intelligence infrastructure while also backing a voluntary commitment intended to prevent data center development from increasing household utility bills.
States confronting particularly rapid development have begun adopting their own approaches.
Texas Gov. Greg Abbott has directed regulators to stop data center projects seeking connections to the state’s primary power grid until the proposals undergo comprehensive audits.
The move reflects concerns about whether new projects could place excessive demands on the grid and affect both reliability and electricity costs.
Pennsylvania Gov. Josh Shapiro has similarly pursued greater oversight of large-scale data center construction as officials weigh the economic benefits of new investment against the additional strain imposed on the state’s power system.
New York Gov. Kathy Hochul has taken a more restrictive approach, imposing a one-year moratorium on new hyperscale data centers.
The varying policies illustrate the broader problem states are confronting as AI companies compete to secure electricity for rapidly expanding computing infrastructure.
Data centers can bring billions of dollars in investment and generate economic activity, but their unusually high and continuous electricity consumption can also accelerate the need for expensive grid expansions that might otherwise have occurred years later.
Areas with clusters of data centers can face particularly significant effects because numerous large facilities may seek power connections within the same geographic region.
The Dallas Fed findings suggest those pressures are no longer merely theoretical. Researchers estimate that existing data center development has already added between 2% and 6% to wholesale electricity prices nationally, while regions experiencing heavier concentrations have seen still greater increases.
The eventual effect on consumers will depend on electricity markets, regulatory decisions, infrastructure investments and how rapidly additional generating capacity can be brought online.
Wholesale power increases also do not necessarily reach consumers immediately because utilities frequently purchase electricity through contracts and operate under regulated rate structures that can delay changes in retail prices.
Nevertheless, as new power plants, substations and transmission infrastructure are constructed to accommodate rising demand, regulators will increasingly have to determine which customers bear those costs.
That debate is likely to become more consequential as artificial intelligence infrastructure expands.
The technology industry’s demand for computing capacity shows little sign of slowing, while electricity grids must simultaneously serve households, traditional businesses, factories and other existing users.
The Dallas Fed research points to a growing tradeoff confronting policymakers: the United States can continue rapidly expanding the physical infrastructure supporting artificial intelligence, but doing so will require substantially more electricity and potentially billions of dollars in additional energy infrastructure.
How those expenses are allocated could determine whether the AI boom becomes increasingly visible on one of the bills Americans have little choice but to pay each month.
Original article
Comments