Even as “abysmal” permitting processes for new power projects prevent grid expansion to meet the massive energy needs of artificial intelligence (AI) companies in the United States, AI technology itself could be deployed to help increase grid efficiency.
New power projects have a hard time connecting to the U.S. grid amid frequent legal fights from environment and community groups, plus layers of regulations and permitting processes, writes Reuters.
The process is “abysmal,” KPMG’s financial management lead for the power sector, Brad Stansberry, told a recent AI and energy sector conference in Washington, DC.
“It’s difficult to get anything built quickly,” agreed Michael Keyser, who heads the National Renewables Cooperative Organization. “So I think it’s a natural outcome of that for why you see data centres and tech companies pursuing their own power purchasing now.”
As new technologies like generative AI increase demand for electricity, tech companies are forced to contract with power producers directly—or build their own supply.
“We always assumed there would be power, but obviously that assumption was woefully inaccurate, so now we’re chasing where power is,” said Phill Lawson-Shanks, chief innovation officer at data centre operator Aligned.
Aligned is looking to solve its power supply problem with small modular reactors, while OpenAI CEO Sam Altman is reportedly looking into nuclear fusion—the energy generating process that powers the sun, but is at best decades away from being used here on earth. [Good luck with that—Ed.]
Experts say Altman’s plans are too far off to deal with the present difficulty of powering AI amid a climate crisis, when all efforts need to focus on reducing emissions from electricity generation, writes CNN.
Meanwhile, AI could help improve grid performance when used with advanced analytics “to enhance grid forecasting, planning, security, orchestration, and more,” finds a new U.S. Department of Energy (DOE) report. It is already being used in this way, but there will be more opportunities and greater need for AI as “managing the complexity of a more dynamic modern grid grows,” DOE says.
Researchers say the impact of AI’s energy use on carbon emissions can’t be assessed by only looking at its direct impacts. The indirect outcomes of AI—which can be positive or negative—also need to be considered.
“Most discussions so far about AI’s environmental consequences have focused on the direct impacts of these computationally intensive technologies—how much energy, water, or other resources they consume and the amount of greenhouse gases they generate,” writes a team of researchers in the journal Nature.
“But the global repercussions of AI applications for society will be much broader, from transforming health care and education to increasing the efficiency of mining, transportation, and agriculture.”
