Key takeaways from the 2025 Experience POWER conference
By Cody Berra, Senior Solution Consultant at UptimeAI & John Velez, Power Generation Account Executive at UptimeAI
With the equivalent of two Californias worth of generation joining the US grid in the next five years, it’s no wonder the industry is feeling some pressure. At Experience POWER, industry leaders gathered in Denver, CO to share how they’re approaching today’s market reality, what challenges they face, and the role AI has to play.
The Industry’s Brutal Reality
Throughout the keynotes and conversations, industry leaders shared sobering data about the pace of load expansion and the readiness of today’s grid to handle it. They spoke candidly about the challenging realities the industry faces: an aging asset base, a rapidly retiring workforce, and a growing share of variable renewable generation that challenges grid stability. But there was also optimism. Mike Caravaggio, VP of Reliability at EPRI, voiced that “This is the most exciting time in the industry in the last 25 years.”
Market statistics presented show the grid nearly doubling over the next decade, but in a way that looks different from past expansions. We’re expanding in such a way that a gigawatt of capacity doesn’t necessarily equal a gigawatt of energy. With most new capacity projects being variable renewable generation, the existing fleet will be tapped to grow that capacity, which is where things get interesting.
Pushing the Potential of Aging Assets
Panelists from Xcel Energy, Platt River Power, and Great River Energy were united by the fact that they’re all dealing with aging assets with a 70-year asset life. These assets, designed for baseload operation, are now being cycled to accommodate renewable generation. Increased utilization combined with frequent start-stop cycles are only adding to the reliability challenges.
The existing natural gas fleet is playing a critical role as a “bridge fuel” for sustaining reliable power generation in the “dark and calm” times when renewables struggle. Long lead times (>3 years according to GE Vernova) on turbine orders solidify that this balance will come from the existing asset base.
Getting more out of current assets will mean driving operational excellence. Improving process efficiency, asset reliability, and optimizing maintenance cycles are some of the levers that will be required. For most utilities, making meaningful gains on these metrics still relies heavily on experienced engineers and operators working with technology.
But Reliance on Expert Workforce is Unsustainable
Another message that reverberated during the sessions was that capacity growth is complicated by the fact that knowledge and experience is exiting the industry and it isn’t being replenished at the same rate. This shrinking workforce is creating a skills void, particularly in senior operations and technical expertise. These are the people who can walk by a pump, hear it, and tell you it’s going to fail in two weeks.
Stephen Martz from Xcel Energy pointed out the hard truth: “Electrical engineering grads want to go work for Intel or NVIDIA. The last thing they want to do is work for their local utility.” Rather than try to build and recruit a workforce that doesn’t exist, “we need to really start looking at how you leverage AI to fill the gap.”
Which is Where UptimeAI Comes In
In a panel session led by our technology partner Yokogawa, we got a chance to share our experience implementing AI solutions to address the reliability and expertise challenges in the power generation industry. Cody drew on his memories of his first months working in a plant environment, and his reliance on the seasoned experts to help him solve problems that spanned siloed data sources and domains. New people walking into a plant today aren’t given that same access to expert resources. But they are given access to exponentially better technology than was available even a decade ago.
While the previous decades’ tools focused on anomaly detection, this created flurries of alerts requiring expert interpretation to sift through false positives, diagnose root causes, and prescribe actions. This time lag between detection and action sometimes spanned hours to weeks depending on the workload of the plant experts. You could try to throw more people at the problem, but it wasn’t scalable.
Taking advantage of modern generative AI has opened a whole new world beyond traditional predictive analytics technologies. UptimeAI is encoding the thought process and workflows of a seasoned plant SME into technology that can reason, like your plant experts, across sources of information and experiences to eliminate the detection to action lag. This translates to successfully mitigated failure events that reduce equipment downtime, save on maintenance costs, and eliminate environmental and personnel safety risk.
When asked what’s working and what’s not in the world of technology projects, Cody shared the following recipe—Don’t look for a shiny tool and think about all the things you might be able to do with it. Start with the specific problem you’re trying to solve, then partner with technology providers with proven success in solving that problem in your industry. Choose your pilots wisely… that’s how you get the momentum you need, to get the funding you need, to bring it all to fruition.
To see firsthand how AI that works like your experts will elevate your asset reliability to meet the rising demand, get a demo.