AI has high energy demands but could also speed up energy production, officials say

Meeting the growing energy demands of data centers and artificial intelligence technologies will require speeding up the often protracted permitting process for deploying clean energy infrastructure, but AI tools can also help accelerate these mandatory efforts, according to federal officials. 

Speaking at a Brookings Institution event on Monday, Emma Nicholson — a senior economic advisor with the Federal Energy Regulatory Commission — said “nothing has come on the scene faster than this new load growth from AI and data centers in computing loads.”

A May report released by Goldman Sachs Research estimated that, as a result of the increasing use of AI technologies, the data center power demand will likely grow 160% by 2030.

This increase in AI-powered data centers — coupled with other electricity demands, such as efforts to onshore manufacturing in the U.S. — means that the slow pace of regulatory steps needed to deploy additional energy infrastructure could fail to keep up with the nation’s power needs. 

“We’ll have to do a lot of streamlining of permitting,” Nicholson said, adding that incorporating more stakeholders into the process is also a major consideration when it comes to expanding the nation’s energy grid. 

Deploying clean energy infrastructure often depends on close collaboration between relevant agencies and state and local officials to meet relevant government regulations and regional ordinances. This multifaceted process can often take years.

Although AI is poised to lead to increasing power consumption moving forward, the tools also represent an opportunity for officials to speed up the permitting process, which could, in turn, help meet growing U.S. energy needs.

President Joe Biden’s October 2023 executive order on the safe, secure and trustworthy development and use of AI directed Energy to develop “models that streamline permitting and environmental reviews while improving environmental and social outcomes.”

During Monday’s event, Helena Fu — director of the Department of Energy’s Office of Critical and Emerging Technologies — said AI-related energy consumption and other increasing power demands on the U.S. energy grid represent an area where “we’re going to need to see some innovation.”

Energy is already investing heavily in AI research and development, Fu noted, in part when it comes to how the technologies can speed up the permitting process. 

The department is spending almost $20 million to test and deploy AI tools that can speed up the permitting process, particularly when it comes to using the capabilities to quickly analyze historical permitting and environmental reviews conducted under the National Environmental Policy Act. The law requires agencies to conduct environmental assessments of major actions, including when it comes to issuing permits for energy infrastructure. 

One of these initiatives, known as VoltAIc, is designed to help officials access the last two decades of NEPA reviews. Fu said the goal is to make the relevant data “AI ready and available and released into the community for academics and industry to build tools around.”

The initiative is still in beta testing, and Fu said Energy is currently working with the Permitting Council to sign memorandums of understanding “to potentially use this.”

Beyond working to speed up the permitting process for new clean energy infrastructure, officials believe that emerging tech capabilities can also help with broader grid deployment and oversight.

In an April report, Energy said that AI “can help lower the amount of new buildout required to reach a 100% clean grid by unlocking underutilized assets, improving coordination and optimization of new infrastructure and by synthesizing the vast and diverse types of technical data needed through foundation models.”