Episode Summary
Executive Summary: The episode argues that U.S. new nuclear is in a genuine but still fragile “groundswell,” driven by AI load growth, hyperscaler demand, capital, and improving policy. Chris Colbert of Elemental Power says the winning model is customer-led, multi-site, technology-agnostic development with utilities and OEMs sharing risk, aiming for early-2030s commercial operation and a repeatable cost-down curve.
Main Topics: U.S. nuclear market momentum and the “groundswell” thesis (Priority: 5/5): Shail Khan frames the current U.S. nuclear boomlet as a groundswell: real momentum from policy, demand, and capital, but not yet guaranteed to become a wave. The discussion centers on whether this moment can translate into actual grid capacity. Hyperscaler demand and offtake are changing project finance (Priority: 5/5): Colbert explains that unlike the first nuclear renaissance, today’s demand is driven by AI/data centers and hyperscalers willing to do more than sign PPAs. They are now expected to help fund development risk upfront, making projects financeable. Technology landscape: multiple OEMs, but no clear winner yet (Priority: 4/5): Elemental is technology-agnostic, waiting for reactor designs to clear licensing, supply-chain, and cost hurdles. Colbert expects several vendors to become commercially viable, but says the market is not ready to pick winners today. Regulatory progress is real, but incomplete (Priority: 4/5): The NRC has sped up reviews, with timelines improving materially, yet no new design has reached an operating license for fuel loading. The panel sees a path toward shorter timelines and lower uncertainty, especially for light-water reactors. Utilities as partners, not developers (Priority: 4/5): Utilities remain essential because they control territory, interconnection, and operations, but many cannot take development risk. Elemental’s model is to absorb development/construction risk and work with utilities and hyperscalers on transfer or operational roles. Cost reduction through repetition, scale, and disciplined execution (Priority: 5/5): Colbert argues nuclear cost declines require repeat builds, multiple units/sites, and careful labor and supply-chain planning. The first project will be expensive, but each subsequent unit should be cheaper if the program is repeatable. Site selection and project-development process for new nuclear (Priority: 4/5): Elemental’s workflow reverses the traditional model: secure customer demand first, then select sites with the right utility, transmission, and seismic characteristics, then pursue an early site permit and construction permit.
Key Arguments: The current U.S. nuclear moment is different from the 2007-era renaissance because AI-driven load growth and hyperscaler balance sheets create real demand and financing power. Corporate buyers must do more than sign PPAs; they need to commit development capital because nuclear development costs are far larger than renewables. Elemental’s technology-agnostic approach is intended to wait for reactors to clear licensing, supply chain, and cost-estimation hurdles before choosing winners. Light-water reactor designs are viewed as lower technology risk because they are proven variants of established technology; non-light-water designs likely arrive later. Regulatory timelines are improving, but nuclear still needs further acceleration to resemble conventional gas plant schedules. Utilities generally cannot take on the development risk required for new nuclear, so developers must bridge that risk gap and then hand off or operate assets with utilities. Repeatability is the main path to lower nuclear costs: multiple units, multiple sites, and standardized execution improve learning curves just like repeated assembly of identical products. Site selection is driven more by the hyperscaler’s preferred utility relationship than by geography alone, provided the site meets seismic and transmission requirements. Transmission access is a critical constraint, and in many cases harder to permit than the nuclear plant itself. The likely commercial rollout for new nuclear in the U.S. is early 2030s, with first-unit costs declining as later units benefit from experience.
Data Points: Transition AI conference date: June 12, 2025 - Latitude Media promotion at the start of the episode Google / Elemental nuclear portfolio: At least 1.8 gigawatts - Elemental’s announced three-site U.S. new nuclear partnership with Google Number of sites in Google partnership: 3 sites - Elemental’s portfolio announcement Hyperscaler development capital expectation: Hundreds of millions of dollars - Colbert says nuclear requires upfront development spending far beyond renewable project development Existing/customer device aggregation: 2.5 million customer devices - EnergyHub ad read describing VPP aggregation capacity Dispatchable capacity from devices: 3.4 gigawatts - EnergyHub ad read describing virtual power plant output Equivalent grid capacity: More than three nuclear reactors - EnergyHub ad read comparing VPP capacity to nuclear reactors Nuclear review timeline improvement: From 4 years to 2 years - Colbert cites NRC review time reductions over his 15 years in the sector Targeted further NRC review timeline: 1 to 1.5 years - Colbert’s hoped-for future review duration Current first-project timeline: 7 to 8 years - Colbert’s estimate for initial new nuclear projects from development to operation Potential improved project timeline: 5 to 6 years - Colbert’s estimate if the industry continues to improve execution and permitting Early deployment target: Early 2030s - When Colbert expects commercial deployment for light-water SMRs/new nuclear Planned growth targets discussed: 1 GW under construction by 2030; 10 GW by 2035; 100 GW by 2040 - Colbert describes Elemental’s broader industry ambition Updated near-term target: 4.2 GW by 2030 - Colbert says the sector has already moved beyond the earlier 1 GW goal Cost target: About $10,000 per kW - Colbert’s expected mature-cost level for new nuclear Example plant size and cost: 600 MW plant ≈ $6 billion - Based on the $10,000/kW cost assumption First-unit cost premium: 15% to 20% higher - Colbert says the first plant will likely cost materially more than later ones Project scale for Google deal: About 600 MW+ per project - Three projects totaling 1.8+ GW Site footprint for smaller modules: About 100 acres or less - Estimated site size for small modular reactor deployments Utility/customer device count in VPP: 170 utilities - EnergyHub ad read about utility participation
Pivotal Quotes: "I think that there's the right confluence of both market demand, investor sentiment, and regulatory and government policy to really make it happen this time." — Chris Colbert: Colbert explains why this nuclear cycle differs from the 2007 renaissance "We went out, find a customer... willing to put in a portion of the development capital." — Chris Colbert: Describes Elemental’s customer-first development model for new nuclear "The first one is going to be more expensive than the second one than the third one." — Chris Colbert: On cost declines through repeat builds and standardized execution
Implications: New nuclear is moving from concept to bankable development, but success depends on repeat builds, regulatory progress, and hyperscaler commitment. Expect early-2030s deployments first, with costs falling only if the industry sustains multi-unit programs.