Data centers aren't a real problem for US power. Douglas J. Arent: Outdated policy is.

The data centers are blamed for the rising cost of electricity in America. The real problem is not that AI and consumers are increasing energy consumption. The real issue is structural and began before the recent infrastructure boom.

The average residential electricity rates increased by 6% in the past year, which is more than double the inflation rate. About one-third of American homes now spend over 5% of their earnings on electricity.

Investor-owned utilities will file the most rate increase requests in 2025. This is their highest level since mid-1980s. There is clearly a problem.

The data centers that are driving the AI race and other industries moving towards electrification can be credited with these increases.

Electricity bills for residents in some states, such as Nebraska, New Mexico, and North Dakota have decreased.

According to two studies conducted by the Columbia University Center on Global Energy Policy, power costs are rising faster than inflation in the Mid-Atlantic region, California, the Northeast, and the Southeast, areas where data centers have been less prevalent.

Why does demand growth lower bills in some areas and raise them in others?

Two words: poor incentives.

BUILT TO SPLEND

Since the 1950s, American utilities have been rewarded for building new infrastructure and not for managing their existing assets. According to Federal Energy Regulatory Commission 'filings,' utilities receive reliable returns on their capital investments, typically between 9% and 10%. Upgrade an existing line or deploy software in place of new infrastructure instead? Answer is not so clear. Customers are liable for the cost of large capital investments, as the incentive is built in. In the past decade, this dynamic has been less viable due to the soaring inflation in the energy sector. Inflation, tight supply chains, and increased tariffs are driving up the cost of transformers that transfer electricity between circuits. The price of wire and cable has risen by 152%. These costs could be reflected in?customer's bills for many decades.

Climate change is another issue. Some utility bills in Florida now include "storm cost recovery surcharges". California bills have increased in the last five year to reduce wildfires. These increases are not an anomaly. These are ongoing, compounding costs that the ratepayers have to absorb.

This?backdrop' was the backdrop against which the data center boom occurred. It did not create a power affordability problem, but it exposed and accelerated an existing one.

Not Keeping Pace

This does not mean that data centers are benign.

Data center power demands will range from 5 megawatts to 200 MW by 2024. This is equivalent to about 200,000 homes. Massive new data campuses, with power demands of 1,000 to 5, 000 MW, have been proposed and are currently under construction. Data centers are expected to use 5% to 15 % of the total U.S. electricty by 2030. This has caused concern across the nation.

Grid upgrades are not free.

In areas where the grid is overloaded, new large loads can increase local costs. This is especially true if utilities pass these expenses on to consumers.

But the answer is not to limit demand. In fact, a broader view of the evidence suggests different solutions.

New electricity demand can lower prices for all when wind and solar energy is available at low cost and where large users pay their fair shares. This outcome is not guaranteed. The grid's ability to connect with low-cost supplies and fairly allocate upgrade costs will determine the outcome. It could be that data centers are required to pay for transmission upgrades.

Lower Costs, Better Rules

The major problem for U.S. grid operator is that infrastructure required to supply higher volumes of energy has not kept up with demand growth. PJM is the biggest grid operator in America, and it serves 13 Mid-Atlantic states. CGEP's studies reveal that there are real solutions available. Innovative uses of existing technology could increase the capacity of transmission lines already in place. Dynamic line ratings, for instance, use real-time weather information to determine the actual capacity of electricity lines, instead of relying solely on static assumptions which often overestimate what is needed. This approach helped a Pennsylvania?utility reduce congestion on monitored?lines by as much as 65%.

Upgrades to the lines themselves are another option. Replace steel-core wires by lighter, stronger carbon core alternatives to nearly double line capacity within months.

CGEP analysis shows that deploying grid-enhancing technology nationwide could result in savings of $180 billion by 2050.

Data centers can be part of the solution. Pilot projects in Arizona and North Carolina have demonstrated that data centers could be designed so as to not draw power from the grid when there is high demand.

These tools are a temporary solution. But more fundamental reforms are needed.

One possibility is to tie executive compensation for utilities to the efficiency of their systems, and not just on how much they build. Modernizing the permitting, interconnection and other processes that slow down new power plants and transmission systems could be a game changer.

Data centers, and other large, energy-hungry infrastructure, could also shoulder a fair portion of the energy costs?they generate, rather than passing these costs on to residential ratepayers.

States that have seen a decrease in electricity prices despite an increase in demand from data centres are not magic. They managed their supply, infrastructure and cost allocation in a sensible way. When these things are not managed properly, prices rise. These things do not need to.

(source: Reuters)