Data centers use significant electricity

Data centers are the fastest-growing electricity consumers in the United States, having consumed between 177 and 192 TWh in 2024, or 4 to 5 percent of the country's electricity. This share could rise between 9 and 17 percent by 2030, according to the Electric Power Research Institute (EPRI).

New alliance promotes flexible energy use

On September 16, Google, Nvidia, and Emerald AI joined forces to launch the AI Energy Management Alliance (AEMA), an initiative aimed at promoting this approach. The alliance includes 18 launch partners, including Anthropic and energy companies such as Constellation, National Grid, NRG, and RWE.

Flexible energy solutions for data centers

The principle of flexibility involves adjusting infrastructure consumption during periods of high demand. Solutions include postponing non-urgent computing tasks, transferring calculations to less strained data centers, and using on-site batteries to store electricity for use during peak times.

Adjusting equipment to save energy

Operators can also adjust the operation of certain equipment, such as cooling systems, to reduce consumption during critical periods. Nvidia is already testing this principle in Virginia on Aurora, a 96 megawatt site built by Digital Realty. The Emerald AI software there shifts the most power-hungry AI calculations in time or to other sites when the network is stressed, then restarts them when demand decreases.

Potential for reducing peak power demand

According to Arin Kaye, technical director at EPRI, data centers surveyed by the institute say they are capable of reducing their peak power demand by 10 to 30 percent depending on the type of installation, and even more for some hyperscalers. These reductions could represent a significant contribution to the grid during high demand periods.

Limitations of current flexibility efforts

However, these reductions are mainly limited to pilot projects or case-by-case agreements. OpenAI, for example, has committed to reducing up to 1 gigawatt of power drawn from the grid during periods of tension, out of the 3.2 gigawatts planned for a future site in Georgia.

Economic benefits of flexibility

By limiting consumption peaks, flexibility could avoid or delay costly investments in electrical infrastructure. A study by the Nicholas Institute at Duke University estimates that this flexibility could save between 40 and 150 billion dollars (approximately 35 to 130 billion euros) in investments over the next decade.

Regulatory support for flexible data centers

These potential savings explain the interest of authorities in this new approach. In October 2025, Energy Secretary Chris Wright asked the Federal Energy Regulatory Commission (FERC) to accelerate the connection of large consumers, with an expedited procedure for those who agree to reduce their power demand. On June 18, FERC ordered the six major regional grid managers to review their connection rules, opening the way for faster connections for flexible installations.

Challenges in implementing flexibility

While the flexibility of data centers can contribute to relieving American electrical networks, its generalization requires technical and economic adaptations. Not all computing tasks can be postponed without consequences. Operators must determine which workloads can be adjusted and up to what level power can be reduced without degrading performance or disrupting services provided to customers.

Investments needed for large-scale flexibility

Large-scale deployment also requires investments in control software, consumption management equipment, and, if needed, storage batteries. For operators to manage this, grid managers and energy providers will have to offer appropriate tariffs, compensation for power reductions, or faster connections. However, flexibility will only be one of the solutions allowing the growth of data centers to be integrated into the electrical system. On the production side, FERC has also required each regional grid manager to explain how it plans to ensure sufficient capacity to supply these large consumers, both existing and new.