Data centers guzzle huge amounts of power—and that might end up supercharging investment in renewable energy, executives argue

The AI boom is driving demand for data centers as these specialized facilities provide the computing power and data storage needed to train and operate AI applications like large language models.

But data centers have garnered criticism for their potential strain on water resources and energy. The International Energy Agency reports that some AI-focused data centers consume as much electricity as 100,000 homes, with the largest facilities under construction potentially using as much as 20 times as much power. 

Yet data center executives argue that demand for their facilities could end up accelerating the shift to renewable energy as companies and governments try to find ways to grow sustainably. 

In the near-term, data centers will need to be built near existing power plants, which are likely powered by fossil fuels, Rangu Salgame, CEO and co-founder of Princeton Digital Group, explained on Tuesday at the Fortune Brainstorm AI Singapore conference. 

But the medium-term, past the next five years, will be “all about energy transition,” he said, citing power sources like solar and natural gas. Nuclear energy might even be a compelling option on a ten-year time horizon. 

Cutting the amount of energy needed to run data centers will also result in large savings, said Tim Rosenfield, CEO of Firmus & SMC Cloud. He agreed that data centers and AI factories can help push renewable energy development, citing the experience of one of his projects in Tasmania.

“An AI factory can be like a new smelter,” he said, likening such projects to other energy-hungry facilities that can drive investment in new power plants. “If we come into a region and we need to find power, we can buy power from new renewable projects.”