Case study: Advanced nuclear (SMRs & Gen IV) — a pilot that failed (and what it taught us)

On November 8, 2023, NuScale Power and the Utah Associated Municipal Power Systems (UAMPS) announced the termination of the Carbon Free Power Project (CFPP)—what was to be the first commercial small modular reactor deployment in the United States. The project's collapse sent shockwaves through the advanced nuclear industry: NuScale's stock plummeted 33% in a single day, erasing over $500 million in market capitalization. What had been positioned as a breakthrough demonstration of SMR viability instead became a $232 million lesson in first-of-a-kind (FOAK) risk management. By 2025, the Institute for Energy Economics and Financial Analysis documented that every completed SMR project globally had experienced cost overruns between 300% and 700%, with construction timelines averaging 12 years rather than the 3-4 years promised. The CFPP failure offers policy professionals and compliance officers critical insights into licensing pathways, FOAK-to-NOAK cost curves, and the supply chain readiness gaps that continue to challenge advanced nuclear deployment.

Why It Matters

The advanced nuclear sector sits at a crossroads. Data centers are projected to consume 8-12% of US electricity by 2030, up from 4% in 2024, driving unprecedented demand for firm, carbon-free power that renewables alone cannot reliably provide. Grid operators increasingly recognize that intermittent solar and wind require complementary baseload sources to maintain the 99.97% reliability standard expected of modern grids. Nuclear power—with capacity factors consistently exceeding 90%—offers the only proven zero-carbon technology capable of providing 24/7 generation at scale.

Yet the economic case for new nuclear construction remains precarious. The CFPP's target power price escalated from $58/MWh in 2020 to $89/MWh by January 2023—a 53% increase that rendered the project uncompetitive against natural gas at $45-50/MWh and utility-scale solar at $32/MWh. Capital cost estimates ballooned from $4.2 billion for a 600 MW plant to $9.3 billion for a scaled-down 462 MW facility, pushing the per-kilowatt cost from $9,964 to $21,561—a 116% increase that fundamentally undermined the SMR value proposition.

For policymakers evaluating advanced nuclear as a decarbonization tool, the CFPP failure illuminates structural challenges that transcend any single project. The subscription model—requiring 80% utility commitment before construction—secured only 26% participation as costs escalated beyond what municipal utilities could absorb. The Department of Energy's $1.4 billion cost-share award, of which approximately $232 million was obligated before cancellation, demonstrated that even substantial federal support cannot overcome fundamental market mismatches. Understanding why this project failed—and what subsequent developments suggest about viable pathways forward—is essential for designing policies that accelerate advanced nuclear deployment without exposing ratepayers or taxpayers to unmanageable risk.

Key Concepts

First-of-a-Kind (FOAK) Premium: The cost differential between initial commercial deployments and mature technology reflects irreducible uncertainties in manufacturing, construction, and regulatory compliance. NuScale invested $1.8 billion solely to achieve NRC design certification—a non-recurring cost that future projects will not bear. However, the CFPP experience revealed that FOAK challenges extend far beyond regulatory expenses. Supply chain immaturity, workforce inexperience, and site-specific engineering adaptations contributed to cost escalation that exceeded initial contingency allocations by factors of 2-3x. Industry projections suggest NOAK (Nth-of-a-Kind) deployments could achieve 40% cost reductions by the 10th unit, but no Western SMR has reached serial production to validate these projections.

Subscription Model Risk Allocation: UAMPS structured the CFPP as a collective purchase agreement among 48 municipal utilities, with project advancement contingent on achieving subscription thresholds at defined "off-ramp" decision points. This structure, designed to protect small utilities from bearing disproportionate risk, ultimately doomed the project. As costs escalated, utilities faced a collective action problem: each member's withdrawal increased per-unit costs for remaining subscribers, triggering additional departures in a death spiral. By cancellation, only 116 MW of the 462 MW target had been subscribed—leaving the project with 25% participation rather than the 80% required for construction financing.

Levelized Cost of Electricity (LCOE): The standard metric for comparing generation technologies on a per-MWh basis incorporates capital costs, fuel expenses, operations and maintenance, and financing charges over the plant's operational lifetime. NuScale's LCOE trajectory from $55/MWh (2016 projection) to $89/MWh (2023 estimate) to $102/MWh (inflation-adjusted 2030 projection) illustrates how SMR economics remain highly sensitive to construction timeline and interest rate assumptions. A 2-year construction delay can eliminate SMR cost advantages entirely due to accumulated financing charges during the construction period.

NRC Design Certification and Site Licensing: The Nuclear Regulatory Commission employs a two-track process: vendors seek Standard Design Approval (SDA) or Design Certification for their reactor technology, while project developers pursue site-specific Combined Operating Licenses (COLs). NuScale achieved SDA for its original 50 MWe module design in September 2020—the first SMR certification in US history—and received approval for the uprated 77 MWe design in May 2025. However, the CFPP's cancellation occurred after NRC certification, demonstrating that regulatory approval is necessary but insufficient for commercial deployment.

What's Working and What Isn't

What's Working

Regulatory Streamlining Through the ADVANCE Act: The 2024 ADVANCE Act reduced NRC hourly review fees from $318 to $148 for "first mover" reactor technologies, substantially decreasing licensing costs for subsequent SMR developers. The legislation also directed NRC to establish risk-informed, performance-based regulatory frameworks better suited to advanced reactor designs than rules originally developed for large light-water reactors. These reforms directly address cost drivers that contributed to NuScale's $1.8 billion pre-construction investment.

Creditworthy Off-Takers Replacing Municipal Utility Consortiums: Post-CFPP SMR development has shifted toward customers with substantially stronger balance sheets. Google's 2024 agreement with Kairos Power for 500 MW of SMR capacity, Amazon's partnership with X-energy, and ongoing data center negotiations signal that hyperscale technology companies—with market capitalizations exceeding $1 trillion and documented willingness to sign 20-year power purchase agreements (PPAs)—can absorb FOAK cost uncertainty that municipal utilities cannot. These arrangements transfer risk to parties capable of bearing it while providing project developers with the revenue certainty needed to secure construction financing.

Experienced Nuclear Operators as Development Partners: Tennessee Valley Authority (TVA) and Ontario Power Generation—utilities with decades of nuclear operating experience—have emerged as preferred SMR customers. TVA's ongoing evaluation of multiple SMR technologies for deployment at Clinch River leverages existing nuclear-qualified workforce, established NRC relationships, and site characterization work completed for previous projects. This institutional capability addresses workforce and supply chain challenges that contributed to CFPP cost escalation.

What Isn't Working

Fixed Civil Construction Costs Despite Modularity Claims: NuScale's VOYGR design requires a large concrete pool structure for reactor containment regardless of how many modules are installed. Scaling the CFPP from 12 modules to 6 modules reduced power output by 50% but decreased capital costs by only 25%, revealing that the "modular" benefits of factory fabrication apply primarily to reactor components rather than site construction. The pool structure, turbine island, and grid interconnection represent fixed costs that limit economies of scope for smaller deployments.

Supply Chain Immaturity for Nuclear-Grade Components: The advanced nuclear sector lacks mature supply chains for specialized components including reactor pressure vessels, steam generators, and safety-grade instrumentation. NuScale's experience demonstrated that suppliers accustomed to one-off orders for large reactors cannot rapidly scale production for SMR quantities. Material cost increases—including 50%+ escalation in carbon steel pricing since 2020—compounded supply chain challenges. Without committed order books spanning multiple projects, manufacturers have little incentive to invest in production capacity expansion.

Negative Learning Curves in Western Nuclear Construction: Historical data from France, the United States, and the United Kingdom reveals that nuclear construction costs have increased rather than decreased with experience—a phenomenon known as the "negative learning curve." Vogtle Units 3 and 4, completed in 2023-2024, cost approximately $35 billion rather than the initial $14 billion estimate. While SMR proponents argue that factory fabrication will break this pattern, no empirical evidence yet supports this claim. The CFPP cancellation precluded any opportunity to validate learning-curve assumptions for NuScale's design.

Key Players

Established Leaders

NuScale Power — The only SMR vendor with NRC design certification, NuScale has pivoted from the failed CFPP to international deployments and domestic data center opportunities. The company's 77 MWe VOYGR module received Standard Design Approval in May 2025, and active projects include Romania's RoPower (462 MW, front-end engineering underway) and ongoing discussions with TVA. Despite the CFPP setback, NuScale's regulatory achievements provide a foundation for future US deployments.

GE Hitachi Nuclear Energy — Developer of the BWRX-300, a 300 MWe boiling water reactor SMR currently undergoing Canadian Nuclear Safety Commission licensing. Ontario Power Generation selected the BWRX-300 for deployment at the Darlington site, with estimated costs of CAD $12-20 billion for initial units. GE Hitachi's existing manufacturing relationships and proven boiling water reactor technology offer advantages over untested designs, though FOAK cost uncertainty remains substantial.

TerraPower — Backed by Bill Gates, TerraPower is developing the Natrium sodium-cooled fast reactor with integrated molten salt energy storage. The company broke ground on its Wyoming demonstration plant in 2024, targeting operation by 2030. The design's ability to provide load-following capability complements variable renewable generation, potentially commanding premium value in markets with high wind and solar penetration.

X-energy — Developer of the Xe-100 high-temperature gas-cooled reactor, X-energy has secured partnerships with Dow Chemical for industrial heat applications and with Amazon for data center power. The design's 80 MWe output and pebble-bed fuel technology enable factory fabrication at scales not achievable with light-water SMRs. X-energy's focus on industrial process heat—where competitors are limited—may provide differentiated market positioning.

Emerging Startups

Kairos Power — Developer of the Hermes fluoride salt-cooled reactor, Kairos achieved NRC construction permit approval in December 2023—the first for an advanced non-light-water reactor in over 50 years. Google's 500 MW commitment provides project bankability that UAMPS could not offer NuScale.

Oklo — Led by former NRC Chair Sam Collins, Oklo is developing compact fast reactors targeting remote and industrial applications. The company's 2024 SPAC merger and subsequent partnerships position it for niche deployments where grid connection is unavailable or prohibitively expensive.

Last Energy — Pursuing standardized 20 MWe micro-reactors for European industrial customers, Last Energy's strategy emphasizes off-grid applications with high electricity prices rather than competing in wholesale power markets.

Radiant Industries — Developing a 1 MWe portable microreactor initially targeting defense applications, Radiant's smaller scale may enable faster iteration and manufacturing learning than larger SMR designs.

Key Investors & Funders

U.S. Department of Energy — The primary funder of advanced nuclear development, DOE has committed over $3.2 billion to SMR and advanced reactor demonstrations through the Advanced Reactor Demonstration Program, cost-share awards, and Inflation Reduction Act implementation. The CFPP experience has informed revised program structures emphasizing experienced customers and fixed-price contract structures.

Breakthrough Energy Ventures — Bill Gates' climate-focused venture fund has invested in TerraPower, alongside strategic capital from SK Group and other partners. BEV's patient capital approach—accepting longer time horizons than traditional venture investors—aligns with nuclear development timelines.

Google (Alphabet) — Beyond the Kairos PPA, Google has committed to achieving 24/7 carbon-free energy across all operations by 2030. This corporate commitment, backed by approximately $150 billion in cash reserves, provides demand-side certainty for advanced nuclear development that policy mandates alone cannot guarantee.

Amazon — Through its Climate Pledge Fund and direct infrastructure investments, Amazon has committed to SMR procurement for data center power. The company's 2024 announcement of investments in X-energy and other nuclear developers signals multi-decade commitment to nuclear offtake.

Examples

1. The NuScale CFPP Collapse: A $9.3 Billion Lesson in Customer Selection

The Carbon Free Power Project's trajectory from announcement to cancellation illustrates how customer-vendor mismatches can doom technically viable projects. UAMPS, a consortium of 48 small municipal utilities serving communities across Utah, New Mexico, Idaho, and Nevada, represented the antithesis of an ideal FOAK nuclear customer. Individual members lacked nuclear operating experience, balance sheet capacity to absorb cost overruns, and the institutional capability to navigate NRC regulatory processes.

The subscription model—requiring utilities to commit based on projected rather than actual costs—created adverse selection dynamics. Early commitments came from utilities most optimistic about SMR economics; as costs escalated, these same utilities faced the choice of accepting dramatically higher rates or exercising contractual off-ramps. By the final decision point, 22 of 27 remaining subscribers had withdrawn, leaving insufficient commitment to secure construction financing.

The implementation lesson is unambiguous: FOAK nuclear projects require customers with (1) nuclear operating experience, (2) balance sheet strength to absorb 2-3x cost overruns, and (3) electricity demand profiles that value capacity factor over variable cost. Municipal utility consortiums satisfy none of these criteria. Future projects—including those in NuScale's current pipeline—have shifted toward industrial off-takers and experienced nuclear utilities that can absorb FOAK risk without passing costs to captive ratepayers.

2. China's HTR-PM: 300% Cost Overrun and 12-Year Timeline Reveal Global FOAK Pattern

The Shidao Bay High-Temperature Gas-Cooled Reactor (HTR-PM), China's first commercial advanced reactor, achieved grid connection in December 2021 after a development timeline spanning nearly two decades. Initial construction began in 2012 with projected costs of approximately $400 million for the 210 MW twin-reactor plant. Actual costs exceeded $1.5 billion—a 300% overrun—with construction requiring 9 years rather than the 4-year estimate.

The HTR-PM experience demonstrates that FOAK challenges transcend Western regulatory environments and market structures. Despite China's state-owned enterprise model, vertically integrated supply chains, and accelerated permitting processes, the project experienced cost escalation comparable to Western nuclear construction. Pebble-bed fuel fabrication proved more challenging than anticipated, component quality issues required rework, and first-of-a-kind engineering changes accumulated throughout construction.

For US policymakers, China's HTR-PM offers both caution and encouragement. The caution: even optimal institutional conditions cannot eliminate FOAK cost growth. The encouragement: China is proceeding with follow-on HTR-PM projects at Shidao Bay and other sites, suggesting that NOAK cost reduction is achievable for committed programs. Whether Western market structures can sustain multi-project commitment remains the critical uncertainty.

3. Argentina's CAREM-25: The Risk of Perpetual FOAK Status

Argentina's CAREM-25 (Central Argentina de Elementos Modulares) represents an alternative failure mode: a project that never achieves commercial operation despite decades of investment. Construction began in 2014 with completion originally projected for 2018. As of 2025, the 25 MW prototype remains incomplete, with cost estimates having escalated 600-700% from initial projections.

CAREM-25 illustrates the consequences of treating FOAK development as a perpetual research project rather than a pathway to commercial deployment. Without firm timelines, committed customers, or fixed budgets, the project absorbed Argentina's nuclear engineering capacity without producing operational technology. Each design revision—justified by incorporating lessons from construction experience—extended timelines and increased costs. The absence of market discipline enabled indefinite deferral of difficult decisions.

The CAREM experience informs US advanced nuclear policy by highlighting the importance of deployment commitment. Projects structured as demonstrations rather than commercial deployments may never transition to NOAK status because the economic and institutional forces driving cost reduction require production-scale operations. The ADVANCE Act's emphasis on demonstrating commercial viability—rather than simply technical feasibility—reflects lessons from international FOAK experience.

Action Checklist

• Evaluate customer creditworthiness before project commitment: Require binding offtake agreements from entities with investment-grade credit ratings and demonstrated ability to absorb 100-200% cost overruns. Municipal utility consortiums and subscription models have proven inadequate for FOAK risk allocation.

• Establish fixed-price engineering, procurement, and construction (EPC) contracts: Shift cost overrun risk to vendors and contractors rather than ratepayers or taxpayers. NuScale's cost-reimbursable structure with UAMPS transferred escalation risk to parties least capable of bearing it.

• Require regulatory milestone achievement before major capital commitment: Stage investments contingent on NRC design certification, site licensing, and construction permit approval. The CFPP committed customer resources before regulatory pathway clarity existed.

• Develop domestic supply chain capacity through multi-project procurement: Coordinate SMR deployments to aggregate demand sufficient for supplier investment in nuclear-grade manufacturing capacity. Single-project procurement cannot justify the capital expenditure required for supply chain development.

• Structure PPA terms to reflect FOAK uncertainty: Incorporate cost adjustment mechanisms, schedule contingencies, and performance guarantees that acknowledge the irreducible uncertainty of first-commercial-unit deployment while protecting off-takers from unlimited exposure.

• Mandate transparent cost reporting for publicly-supported projects: Require quarterly disclosure of actual versus projected costs, schedule status, and risk register evolution. The CFPP's cost escalation was not publicly disclosed until projections reached levels that triggered subscriber withdrawal.

FAQ

Q: Why did the NuScale CFPP fail if the reactor design was NRC-certified?

A: NRC design certification addresses safety and technical requirements but does not guarantee commercial viability. The CFPP failed because projected electricity costs ($89/MWh by 2023) exceeded what subscribing utilities could justify to their ratepayers, particularly given competing options like natural gas ($45-50/MWh) and renewables ($30-35/MWh). Additionally, the subscription model concentrated risk on small municipal utilities without the financial capacity to absorb cost overruns. The project would have required $9.3 billion in construction capital, but only 26% of planned capacity was subscribed—far below the 80% threshold needed for financing. Regulatory approval is necessary but insufficient for commercial nuclear deployment; market conditions, customer capability, and risk allocation structures are equally determinative.

Q: What is the realistic FOAK-to-NOAK cost reduction potential for SMRs?

A: Industry projections suggest 40% cost reduction by the 10th unit, based on manufacturing learning curves observed in other capital-intensive industries. However, no Western SMR has achieved serial production to validate these projections empirically. Historical nuclear construction in France, the US, and UK has exhibited negative learning curves—costs increased with experience due to evolving regulatory requirements, quality issues, and construction management challenges. SMR proponents argue that factory fabrication fundamentally changes cost dynamics, but this remains unproven. Conservative assumptions for policy planning should anticipate NOAK costs 20-30% below FOAK rather than the 40% reduction industry claims, with substantial uncertainty in either direction.

Q: Can advanced nuclear compete with renewables-plus-storage on cost?

A: Not on simple LCOE comparisons under current conditions. Utility-scale solar ($32/MWh), onshore wind ($30/MWh), and solar-plus-storage ($45/MWh) undercut SMR projections ($89-102/MWh for FOAK, $50-60/MWh for optimistic NOAK). However, this comparison understates nuclear's value in several respects. Nuclear provides firm, dispatchable capacity with 90%+ capacity factors, while solar (25-30%) and wind (35-45%) require backup generation or storage for grid reliability. For applications requiring 24/7 carbon-free power—data centers, industrial facilities, or grids targeting net-zero emissions—the relevant comparison is SMR costs versus renewable-plus-storage portfolios designed to achieve equivalent reliability. At 99%+ reliability requirements, storage costs escalate substantially, potentially restoring nuclear competitiveness. The economic case for SMRs depends heavily on how grid reliability and carbon-free requirements are valued.

Q: What policy interventions could accelerate viable SMR deployment?

A: The CFPP experience suggests several evidence-based policy directions. First, concentrate FOAK support on projects with creditworthy off-takers rather than distributed subscription models; DOE's revised Advanced Reactor Demonstration Program structure reflects this lesson. Second, extend production tax credits (currently $15-30/MWh under IRA) to provide revenue certainty during the FOAK-to-NOAK transition. Third, establish supply chain development programs that aggregate demand across multiple projects to justify manufacturer investment in nuclear-grade production capacity. Fourth, create risk-sharing mechanisms (loan guarantees, cost-overrun insurance) that transfer construction risk from ratepayers to entities better positioned to bear it. Fifth, maintain regulatory momentum by fully funding NRC advanced reactor review capacity under ADVANCE Act reforms. Critically, policy should recognize that FOAK nuclear is fundamentally different from NOAK deployment and structure support accordingly.

Sources

• Clean Air Task Force. (2023). "Lessons Learned from the Recently Cancelled NuScale-UAMPS Project." CATF Analysis Brief, November 2023.

• Institute for Energy Economics and Financial Analysis. (2024). "Small Modular Reactors: Still Too Expensive, Too Slow, and Too Risky." IEEFA Report, May 2024.

• U.S. Nuclear Regulatory Commission. (2025). "NuScale Power's Small Modular Reactor Achieves Standard Design Approval for 77 MWe." NRC Press Release, May 29, 2025.

• U.S. Department of Energy Office of Nuclear Energy. (2024). "Advanced Reactor Demonstration Program: Lessons Learned and Path Forward." DOE-NE Report.

• Utility Dive. (2023). "NuScale, UAMPS Terminate Small Modular Reactor Project in Idaho." November 8, 2023.

• Nøland, J.K., et al. (2024). "Cost Projections of Small Modular Reactors: A Comparative Analysis." IAEA Conference Paper, International Atomic Energy Agency.

• MIT Center for Advanced Nuclear Energy Systems. (2024). "Advanced Nuclear Power Program: Total Cost Projection for Next-Generation Reactors." MIT Technical Report.

• World Nuclear News. (2023). "Idaho SMR Project Terminated." November 9, 2023.