Leading companies in the nation’s electric power system are looking to artificial intelligence to help find ways to meet growing energy demand and strengthen the grid while reducing greenhouse gas emissions.
Under a consortium announced Thursday, electric utility and power companies from around the world will partner with technology companies to develop open AI models tailored to address the challenges the power sector faces.
“What they’re really looking at is how can we have safer, more affordable, more reliable and resilient energy systems,” Jeremy Renshaw, executive director of AI and quantum for the Electric Power Research Institute, told Newsweek. “AI can help with all of these.”
EPRI, an independent, nonprofit energy research and development organization, helped bring together leaders from the energy and tech sectors to form the consortium. Participants include most of the major U.S. power suppliers and tech giants such as Microsoft, Oracle and NVIDIA.
The consortium was announced at NVIDIA’s Global Technology Conference, an annual event for AI developers in San Jose, California. At last year’s conference, EPRI presented a large language model that had been trained on the institute’s data, a sort of ChatGPT for utility companies. NVIDIA Senior Managing Director for the Global Energy Industry Marc Spieler said the consortium will develop and deploy the next version of that model by drawing on the massive amount of information the power companies have.
“Together, we can combine the data and train a domain-specific industry model that can help accelerate the industry,” Spieler told Newsweek.
Collectively, the power companies and EPRI have reams of data from decades of studies, equipment testing and systems evaluations. But Spieler said they have not been able to put that collective information to use in a cost-effective way.
“We’ll have a lot of that information built into models to help them get access to the answers, not just access to the data,” Spieler said.
Power companies face several daunting challenges in the coming years. After a long period of relatively stable demand, energy analysts say, the need for electricity is surging both in the U.S. and around the world.
Electric grid systems are vulnerable to more disruptions from severe weather and wildfires made more frequent and severe in part due to climate change. Power companies are also working to decarbonize and add more renewable energy.
Clean power from wind, solar and battery storage accounted for the bulk of new electricity generation sources added to the U.S. grid last year. But making those intermittent renewable energy sources work in a complex power system presents its own set of problems, and many power developers face long waits to connect to the grid.
EPRI’s Renshaw said AI might help with each of those problems.
“One area that could be particularly useful is looking at matching generation forecasting with demand forecasting,” he said. Making the right amount of energy at the right time could reduce the need for some of the dirtier power sources on the grid.
Advanced weather forecasting combined with electricity generation forecasting could better sync renewable energy from wind and sun with times of need.
“Then we could reduce some of that older, less efficient and more carbon-intensive energy generation in favor of cleaner technologies,” Renshaw said.
AI can also play a role in speeding up the permitting process for new power sources and transmission lines to help meet rising demand, including the demand for electricity required to run AI.
The explosive growth of data centers to train and operate AI models is contributing to the growth in power demand. By the end of this decade, some studies project, data centers could use up to 12 percent of all the electricity the U.S. produces.
There’s a bit of irony in using power-hungry AI to solve power problems. NVIDIA’s Spieler said the company is making progress on that front as well, increasing the efficiency of the company’s GPU chips at the heart of many AI computing operations.
“Our goal is to continue to drive down the energy required to do the same amount of work,” Spieler said.
Newsweek will highlight some of the best applications of AI in sustainability, health, finance and a wide variety of sectors in the first AI Impact Awards. The awards will recognize innovative AI solutions that solve critical issues. Entries are open until April 25, and finalists and winners will be announced in late May ahead of the AI Impact Summit in June.
