Artificial Intelligence helps fuel new energy sources

"When you have increased demand and inadequate supply, costs are going to go up. And that's what we're experiencing right now," Exelon CEO Calvin Butler said. 

In 2024, U.S. data centers used more than 4% of total U.S. electricity consumption according to the International Energy Agency. That equates to as much electricity as the entire nation of Pakistan uses annually. U.S. Data Center consumption is expected to grow by 133% by the end of the decade, using as much power as the entire country of France. 

"Our growth is unprecedented in the last several decades. So, with the data center advent and the technology coming, we've been forced to serve that load, which is our responsibility," Butler said. "But what we also have to do is build new generation supply, which is not keeping up with the load that is coming on. And that's the crunch that we're in right now."

Commonwealth Edison is asking regulators for a $15.3 billion 4-year grid update to meet the growing demand. The U.S. overall has increased its grid capacity by more than 15% over the last decade, but many utility companies and energy producers say it is not enough. 

Commonwealth Fusion Systems is working to add a new form of nuclear energy to the grid - fusion. It has the same reliable benefits of standard nuclear energy already in use, but does not produce long-lived radioactive waste and carries fewer risks. 

"In fusion there's no chain reaction. The result is helium which is safe and inert and you don't use it to make anything related to weapons," Mumgaard said. 

Commonwealth Fusion Systems says Artificial Intelligence is helping bring fusion energy closer to being a new resource. 

"Building and designing these complex machines and manipulating this complex data matter of plasma are all things that we're still learning and we're still figuring out how to do," Mumgaard said. "And that's an area where we've been able to accelerate using A.I."

Other under-utilized energy sources could soon get a big boost thanks to A.I. Geothermal energy is a small part of the electric grid, because of the high drilling costs and low confidence in where to place infrastructure. 

"If you could drill the perfect geothermal well every single time, like you pick the right spot, you design the right well, you drill the 5,000, 8,000 feet, you hit 400F degree temperatures, that's incredibly productive," Zanskar Co-founder Joel Edwards said. "If you could do that every single time over and over and again, geothermal power is the cheapest source of power period." 

"If we could just get more precise in where we go to find the things and then how we drill into the things, geothermal absolutely has the cost curve to come down," Edwards said. "And that's sort of what we're running towards, with A.I. sort of giving us the boost, giving us an edge to do that."

"It's critical, and we've been raising that alarm for years now, and I use the analogy that you're driving a car and your check engine light is on, but you keep driving it, hoping that you'll keep getting there and keep going, but when it breaks down, you're going to have a significantly higher cost," Butler said. "We have to pay attention to what's going on, and this winter storm - Winter Storm Fern - is indicative of what's coming."