The electricity sector faces a timing problem that’s becoming impossible to ignore. Data centers, artificial intelligence (AI) deployment, industrial reshoring, and broader electrification are driving load growth at rates not seen in decades—and much of that new demand wants carbon-free, firm power. Nuclear checks those boxes. But can the industry deliver capacity fast enough?
POWER spoke with two EPRI vice presidents to get a clearer picture: Steve Swilley, vice president of Nuclear and Chief Nuclear Officer, and Steve Chengelis, vice president of Energy Supply, Nuclear Development, and Fusion. Their message was direct: optionality is not an option. Meeting demand will require simultaneous action on multiple fronts—squeezing more from existing plants, advancing new builds, and solving the execution challenges that have historically plagued nuclear projects.
The Existing Fleet: Nine Reactors’ Worth of Untapped Capacity
The fastest path to new nuclear capacity doesn’t involve breaking ground. It involves getting more from plants already operating.
“Even with accelerated licensing, large new reactors and first-of-a-kind SMRs [small modular reactors] largely extend into the early-to-mid-2030s,” Swilley said. That timeline doesn’t match near-term load growth. What does match it: uprates, outage optimization, long-term operation strategies, and restarts of recently shutdown plants. “EPRI research indicates these measures could deliver the equivalent of roughly nine large reactors of additional U.S. capacity within a few years,” Swilley added.
Historically, U.S. uprates have delivered about 15% increases per uprated unit, totaling roughly 8 GW of added capacity, Swilley explained. EPRI’s current assessment suggests an additional 5 GW to 8 GW of potential remains across the fleet, with dozens of units having site-specific opportunities. The most promising candidates are plants pursuing modernization as part of long-term operation strategies—particularly those upgrading digital instrumentation, improving turbines, or recovering margins through better analytics.
“Achieving the upper end of that range requires coordinated investment in plant modernization, regulatory engagement, and supply chain readiness,” Swilley noted.
Achieving 80-Year Plant Lives
License extensions are enabling plants to operate far longer than originally envisioned, with some now targeting 80-year lifespans. EPRI’s research shows no generic technical barriers to reaching that mark, Swilley said, but extended operation demands deeper understanding of how materials age under decades of stress.
“Key focus areas include reactor pressure vessel embrittlement at high neutron fluence, irradiation-assisted stress corrosion cracking in internals, long-term concrete performance, and electrical cable degradation,” he explained. EPRI is advancing work on nondestructive evaluation techniques, improved surveillance data, and better aging management models.
“The goal is not just license renewal, but sustained or improved reliability and safety over extended lifetimes,” said Swilley.
SMRs: A Changed Business Case
The economic argument for SMRs looks different than it did five years ago—though not necessarily for the reasons the industry once anticipated.
“The business case for SMRs has changed meaningfully, driven less by cost reductions and more by sharply rising demand for firm, reliable capacity,” Chengelis said. Early SMR discussions centered on learning curves and standardization driving down costs. Today, the demand signal itself has become the compelling factor.
Cost inflation remains real, but Chengelis pointed to a different variable as the true differentiator: execution certainty. EPRI’s research—through project development and execution guidance, its Advanced Reactor Roadmap, and work supporting engineering and construction innovation—is focused squarely on improving delivery. “Projects that demonstrate on-time, on-budget performance will ultimately determine the scalability and competitiveness of SMRs,” Chengelis predicted.
The HALEU Question
Advanced reactor deployment hinges partly on fuel availability, specifically high-assay low-enriched uranium (HALEU). “Progress is real but timing remains tight,” Chengelis reported.
Federal investments have helped restart enrichment and deconversion activities in the U.S., but sustained demand signals are needed to make HALEU production commercially durable. EPRI’s work with industry and the Department of Energy suggests early SMR deployments can likely be supported, but scaling beyond first movers requires firm project commitments.
“Fuel readiness is less a technical question than one of coordinated market timing,” Chengelis said.
Supply Chain Deep Cut: The Lithium-7 Challenge
One supply chain vulnerability getting attention involves lithium-7 (Li-7), a material many people outside the nuclear industry may never have heard of. Swilley explained that Li-7 is used in pressurized water reactors to control coolant chemistry and reduce corrosion. It also appears in Generation IV molten salt reactor designs as a coolant or fuel solvent, and has applications in fusion for tritium breeding.
“As global demand for lithium grows across multiple industries, supply concentration and cost volatility become concerns,” Swilley said. “EPRI is evaluating alternative chemistries, such as potassium-based options, that could reduce reliance on Li-7 without compromising materials performance.”
Digital Tools Gaining Traction
Digital twins and AI-driven predictive maintenance are moving from pilot projects toward selective deployment across the nuclear fleet, Swilley suggested. “Utilities are deploying digital representations where the value proposition is strongest,” he said. Those include asset health monitoring, outage planning, chemistry control, and operator support. For new reactor projects, digital twins are supporting design, construction, and materials analysis. “EPRI’s role involves helping to standardize data practices, validate models, and ensure cyber and regulatory considerations are addressed,” Swilley explained.
On the predictive maintenance front, the strongest results come from focused applications rather than enterprise-wide AI rollouts. Anomaly detection for rotating equipment, inspection interval optimization, and condition-based maintenance using fleet-wide data are showing measurable returns.
EPRI’s Preventive Maintenance Basis Database enables anonymized learning across plants, supporting earlier detection of degradation and helping avoid forced outages. AI-enabled diagnostic tools are compressing analysis that once took hours or days into minutes by correlating sensor data, historical performance, and operating conditions.
“These tools are already delivering measurable cost and risk reductions,” Swilley reported. “In these cases, AI acts as a decision-support layer for engineers and maintenance teams, improving confidence and response speed rather than replacing human judgment,” he said.
Fusion: Signs of Progress
The prospects for commercial fusion projects are growing. EPRI is actively engaged with developers representing multiple physics approaches in the U.S. and internationally, Chengelis noted.
“Key signals to watch for include demonstrating energy gain; developing long-duration, high-stability plasma control; and establishing pilot plants,” said Chengelis. “Progress in these areas would indicate that fusion is transitioning from laboratory success toward buildable, operable power plants.”
EPRI’s work with private developers focuses on materials research, safety assessment frameworks, tritium safety, and defining owner-operator expectations. “EPRI’s role is to convene the diverse system of global fusion technology stakeholders to reduce shared technical and regulatory risks that will ultimately determine whether fusion can be deployed on electric grids,” Chengelis said.
The Bottleneck That Matters Most
Asked to identify the single biggest constraint on nuclear expansion over the next decade, Chengelis noted two recurring challenges. “The most consistent bottlenecks EPRI is seeing and addressing with stakeholders worldwide are project financing and supply chain constrictions, both direct influences on execution certainty overall,” he said.
The insight cuts to the heart of the nuclear industry’s most difficult hurdles. Workforce gaps, supply chain limits, licensing timelines, and financing difficulties all become more manageable when projects demonstrate predictable delivery. When schedules slip or costs escalate, every other constraint tightens.
“For both large reactors and SMRs, improving project execution—through standardization, supply chain maturity, and experienced labor—is a clear lever to enable nuclear growth this decade,” Chengelis concluded.
The technology works. The demand is there. The question is whether the industry can deliver projects that prove nuclear belongs in the energy mix at scale.
—Aaron Larson is POWER’s executive editor.
