The growing need for reliable and flexible energy to meet the demands of data centers should drive investments in storage and in stable sources such as nuclear power.
According to a new report from the IEA (International Energy Agency), supply contracts between data center operators and small nuclear reactor projects jumped from 25 GW in 2024 to 45 GW in 2025.
The movement indicates that advances in artificial intelligence could accelerate the commercialization of new energy technologies.
Given the limitations of grid connectivity, developers have also been investing in on-site natural gas power plants, especially in the United States, as a way to ensure security of supply and reduce operational risks.
One of the main challenges for data centers is dealing with rapid load fluctuations. In this context, on-premises energy storage emerges as a strategic technology for the next generation of AI-focused facilities.
With adequate incentives, the IEA He believes that these systems can even act as flexibility assets for the electrical system.
Demand for electricity from data centers grew by 17% in 2025, well above the global expansion of 3%. Meanwhile, capital investments by five major technology companies exceeded US$400 billion during that period and are expected to increase by another 75% in 2026.
The report also points out that, although energy consumption per AI task is falling rapidly, the expansion of the technology's use more than compensates for this efficiency gain. Total electricity consumption by data centers is expected to double by 2030, while those dedicated to AI are projected to triple their demand.
Data from the agency indicates that a Google search consumes about 0,3 Wh of electricity, while a request to ChatGPT requires approximately 2,9 Wh – almost ten times more.
“We now see that while AI still consumes energy, it is also driving innovative solutions such as next-generation nuclear reactors, flexible data centers, and long-duration energy storage,” said Fatih Birol, the agency’s executive director.
Structural bottlenecks
The expansion of data centers poses significant challenges to electrical infrastructure, especially due to the concentration of large loads and the rapid growth of these projects. In many cases, this requires new generation assets and grid reinforcements.
Furthermore, the implementation of these projects faces physical bottlenecks that limit the pace of expansion. Supply chains for critical equipment, such as gas turbines, transformers, advanced chips, and IT components, have become more constrained in the last year.
The increase in the number of projects also puts pressure on the planning of connections to the electrical grid, contributing to delays and uncertainties. Therefore, the strategy of on-site generation is growing, reducing dependence on the grid.
“The IEA was one of the first to recognize that there is no AI without energy – and that countries that offer secure, affordable and rapid access to electricity will be one step ahead,” says Birol.
You might also be interested in!
Data centers, water, and energy: the tipping point of the AI era.
all the content of Canal Solar is protected by copyright law, and partial or total reproduction of this site in any medium is expressly prohibited. If you are interested in collaborating or reusing part of our material, please contact us by email: redacao@canalsolar.com.br.
