In recent years, fusion power has been championed as the ultimate clean energy solution, promising a virtually limitless and carbon-free power source. However, the economic feasibility of fusion energy technology has remained a contentious issue, primarily due to uncertainties surrounding its cost trajectory. A groundbreaking study, published in Nature Energy, now challenges the optimistic assumptions about cost reductions in fusion power plants (FPPs), revealing that experience rates (ERs)—a key metric for forecasting cost reductions—have been consistently overestimated for this transformative technology.
Experience rates quantify how much the cost of a technology decreases each time its cumulative production doubles. Traditionally, ER estimations have relied on extensive historical cost data, accessible for mature technologies like solar panels, wind turbines, and lithium-ion batteries. Fusion energy, being nascent, lacks this extensive cost history, posing a significant barrier to applying conventional ER methodologies. To bridge this gap, the research team introduced an innovative framework by empirically linking the technological characteristics of fusion power plants with ERs observed in analogous energy technologies. This approach allowed them to estimate realistic ER ranges for fusion energy, despite its early developmental stage.
To robustly anchor their projections, the authors conducted 28 semi-structured interviews with eminent experts in both magnetic confinement fusion (MCF) and laser-based inertial confinement fusion (LFE). These industry veterans, spanning public research institutions and private fusion ventures, provided nuanced insights into the inherent complexity and customization needs of FPP designs. The interviews were meticulously structured to minimize bias, utilizing well-defined reference technologies for comparison—solar panels and conventional nuclear fission plants were rated as benchmarks of simplicity and complexity, respectively. This methodological rigor ensured consistent and reliable expert assessments across the board.
The study’s unique survey design included askings experts to rate FPPs on a scale from 1 to 7 regarding design complexity and customization necessity. Solar panels, characterized by relative standardization and simplicity, were set to a score of 2, whereas nuclear fission plants, known for their complex, heavily customized nature, registered a 6. Against this backdrop, fusion plants demonstrated characteristics closer to the higher end of complexity and customization, foreshadowing potentially slower cost descent trajectories. The researchers also explored unit size implications, a factor often intertwined with the technology’s modularity and scalability, which influences learning effects and economies of scale.
The interview transcripts underwent rigorous qualitative coding to extract relevant arguments about the technological traits impacting fusion power’s cost dynamics. These qualitative insights were then integrated with quantitative formulas used to model the evolution of capital expenditure (CAPEX) as a function of cumulative deployment. By applying a generalized cost-learning equation, the team derived experience curves explicitly tailored to fusion energy, positioning them alongside empirically-validated curves for onshore wind, photovoltaic modules, lithium-ion batteries, and nuclear fission.
The familiar learning curve formula they used expresses CAPEX at time t as the initial CAPEX multiplied by the ratio of cumulative deployment at t over initial deployment, raised to the power of the logarithm base 2 of one minus the ER. This formulation elegantly captures the intuitive notion that costs tend to fall more sharply with each doubling of total installed capacity, but the exact slope—i.e., the ER—is technology-dependent and shaped by intrinsic factors such as design complexity and market maturity.
In stark contrast to the rapid historical cost declines witnessed in solar PV—whose ERs often exceed 20%—fusion experience rates, as concluded in this study, likely hover near or below 5%. This suggests that fusion power plants might follow a very gradual learning trajectory, implying that economically competitive commercial fusion may take longer to materialize than some previous models suggested. The findings align fusion energy’s cost evolution more closely with that of nuclear fission, a domain historically plagued by cost overruns and modest economies of scale gains.
The authors further contextualized their findings by leveraging comprehensive databases of historical costs and deployments for other energy technologies. They fitted experience curves for onshore wind, solar PV, and lithium-ion batteries, utilizing published starting costs, cumulative deployment figures, and ER values from authoritative sources, ensuring their fusion projections were anchored in well-understood precedents. For nuclear fission, they applied a log-log linear regression to overnight construction costs as a function of cumulative capacity, further validating the similarity in experience dynamics between fusion and fission plants.
Crucially, all financial figures were inflation-adjusted to 2024 US dollars, maintaining consistency and comparability across diverse datasets. This meticulous approach underscores the study’s commitment to robust, real-world relevance. Its findings invite recalibrated expectations among policymakers, investors, and stakeholders counting on fusion power to drive rapid decarbonization.
The implications of this reassessment extend beyond fusion. They underscore the significance of incorporating detailed technological characteristics into ER estimation frameworks, particularly for emergent technologies lacking extensive cost histories. By doing so, analysts can avoid undue optimism or pessimism, making more reliable projections that guide research funding, industrial strategy, and market development more effectively.
The study builds on and extends prior research linking granular technology traits to learning rates. Through a marriage of qualitative expert elicitation and quantitative modeling, it ushers in a nuanced lens on fusion’s challenging innovation pathway. While the promise of fusion remains undiminished, this research importantly tempers expectations with sober assessments of the inherent complexities and customization demands that may slow down cost reductions.
Moreover, the research illuminates the intricate interplay between technology design complexity, need for customization, and unit size—variables that profoundly influence how quickly a technology can achieve economies of learning. For fusion power, whose systems are bespoke and multifaceted by necessity, these factors collectively shape a less steep cost-learning curve. The insights might encourage fusion developers to strategically pursue modularity and standardization to accelerate cost declines.
Experts participating in the study emphasized that fusion’s relative immaturity and unique engineering challenges differ markedly from renewables, whose components can be mass-produced with well-understood processes. The bespoke nature of fusion reactors—requiring customized magnetic or laser confinement systems, high-grade materials, and precision control mechanisms—pose formidable challenges to realizing rapid experience-driven cost reductions seen in simpler, standardized technologies.
Conducted between August 2024 and March 2025, the interviews provided a temporal snapshot of contemporary expert sentiment across diverse fusion research and industrial landscapes. This temporal grounding enhances the study’s relevance, capturing the evolving fusion innovation ecosystem as it edges closer to pilot demonstrations and potential commercialization.
Overall, this pioneering research invites a recalibration of fusion power’s projected cost trajectory by anchoring ER estimations to concrete technological realities rather than hopeful historical analogies. While commercial fusion remains a pivotal goal for a sustainable energy future, stakeholders must reckon with a learning curve that may unfold more slowly, demanding sustained long-term support and pragmatic planning strategies.
Such sober, empirically grounded assessments are vital for aligning expectations, optimizing resource allocation, and charting practical pathways to a decarbonized energy system where fusion plays a meaningful, albeit gradually unfolding, role. The study’s mixed-method approach, blending expert qualitative wisdom with rigorous quantitative experience curve modeling, sets a robust precedent for evaluating other emerging energy technologies lacking mature commercial track records.
By illuminating the nuanced technological determinants underpinning fusion’s economic evolution, this research moves the discourse beyond simplistic cost extrapolations. It calls for an informed appreciation of fusion power’s engineering intricacies and market challenges, signaling that the road to affordable fusion may be longer and more complex than previously envisaged, but no less essential for humanity’s energy future.
Article References:
Tang, L., Noll, B., Panda, A. et al. Fusion power experience rates are overestimated. Nat Energy (2026). https://doi.org/10.1038/s41560-026-02023-8
DOI: https://doi.org/10.1038/s41560-026-02023-8
Image Credits: AI Generated
