Main

International climate goals to limit anthropogenic warming to 2 °C or less aim to avoid severe climate impacts on natural and human systems in the coming decades1,2,3,4,5,6,7,8,9,10, but such goals entail a rapid reduction in carbon dioxide (CO2) emissions by transitioning away from fossil sources of energy and fossil-fuel-burning power plants in particular11,12. Policies aligned with international climate targets will effectively impose higher costs on emitting infrastructure13,14, leaving the owners of fossil-fuel-burning power plants to adapt and transition away from high-carbon-emitting technologies, absorb substantial financial losses or pass related costs on to consumers. However, the magnitude and distribution of company-level financial risks in the global power sector have not been systematically evaluated, even though understanding such risks is crucial for accurately anticipating broader economic implications and allowing companies, investors and policy-makers to plan for and manage projected financial losses15,16,17,18. In this study, we assess stranded power-plant assets at the company level and by region with the hope of informing financial investors and energy-related decision-makers.

Previous research has shown that achieving ambitious climate goals requires the early retirement of fossil-fuel-burning power plants2,11,19,20 in the absence of retrofitting these plants with carbon capture technology, resulting in substantial forgone revenues. Other studies have estimated potential economic losses across the power sector13,14,21,22,23 or quantified stranded power-plant assets on a national level14,24, but most have not quantified the financial exposure of individual companies. For instance, ref. 24 calculated the proportion of operating fossil-fired electricity-generating capacity that exceeds climate-compatible levels under a 2 °C scenario and estimated stranded assets based on unrecovered capital costs from premature decommissioning. Similarly, ref. 25 quantified stranded assets using overnight capital costs and the unfulfilled portion of the expected lifetime of a coal plant under a 1.5 °C pathway. Meanwhile, ref. 14 estimated stranded assets in oil and gas production as the net present value of foregone profits under climate policy scenarios and aggregated from the asset level to firms and countries. In contrast, we evaluate foregone operating income of existing fossil-fuel-burning generators under carbon pricing estimates associated with lower climate policy scenarios, aggregate these losses to the company and identify and name companies with the largest financial exposure. While financial institutions, investors and policy-makers are generally aware of the financial risks associated with climate policies26,27, the scale and distribution of economic exposure among companies with emission-intensive power-plant assets is not clear. Therefore, this study assesses the financial implications of carbon pricing on fossil-fuel-burning power plants and reports the potential economic losses incurred by individual companies.

An asset is stranded when changes in technology or policy sharply reduce its value. Here we estimate differences in operating income of 16,438 fossil-fired power generators assuming either no carbon price or carbon prices consistent with lower warming targets (1.5 °C, 2.0 °C, 2.2 °C and 2.6 °C) and aggregate these losses on a company and country level. To isolate the financial impact of carbon pricing on generator revenues, we hold electricity generation constant at historical levels across climate policy scenarios and allow generators to pass on 90% of the added carbon costs to the rate-paying consumers. Our default model uses a geographically uniform discount rate, consistent with the parameters widely used by models that generate the carbon prices listed in the shared socioeconomic pathways database28,29,30 (Supplementary Tables 1–4). Further details of our approach are described in Methods. It should be noted that, although our model represents climate mitigation policy in terms of carbon prices, such prices may reflect a multitude of specific policies that are more regionally scaled and not fully captured by our model. Thus, our results are not intended as financial projections for individual companies but rather as a relative indicator of which companies have the highest financial risks under climate policies aligned with international targets if governments do not implement compensatory policies for asset owners.

Results

Figure 1 illustrates the net present value of the added operating costs and foregone operating income from existing fossil-fuel-burning power plants worldwide under carbon pricing scenarios consistent with limiting global mean temperature increases to around 2 °C, assuming each plant continues to operate under its historical conditions into the future. Under these scenarios, the total added costs to fossil fuel power-plant operators are US$18.9 trillion (US$16–21.7 trillion), US$5.4 trillion (US$4.6–6.1 trillion), US$3.7 trillion (US$3.2–4.2 trillion) and US$1.7 trillion (US$1.5–1.9 trillion) based on the median carbon price for each climate policy trajectory, respectively (Fig. 1). Coal-burning power-plant operators would incur 75% of the total costs under a 1.5 °C climate mitigation scenario (Fig. 1a). More than half these costs are concentrated among power-generating companies in China and a substantial share of the costs are borne by companies in India (9.6%), the USA (7.3%) and Indonesia (2.0%) (Fig. 1b). Added costs are highly concentrated amongst countries and companies: 10 companies would be responsible for 24.0% of the total, 90 other companies would bear 33.7%, with the remainder of these costs being divided among 4,851 companies (Fig. 1c). In the absence of strict regulation, power companies are likely to pass through a large share of carbon-related operating cost increases to rate payers, thereby reducing their own financial exposure. However, the portion of costs not passed through by the utility operator would ultimately be absorbed by the company as a loss to operating income (stranded assets). Consequently, the magnitude of financial losses to companies will not only depend on the stringency of climate policy, but also on the proportion of costs passed through to electricity rate payers and the discount rates used to value foregone revenues. We find that stranded assets total US$1.9 trillion, US$0.5 trillion, US$0.4 trillion and US$0.16 trillion under scenarios in which power-generating companies are permitted to pass through 90% of their increased operating costs associated with carbon prices consistent with 1.5 °C, 2 °C, 2.2 °C and 2.6 °C climate targets, respectively, while discount rates are held constant across nations (Fig. 1d). However, revenue losses to fossil power generators are highly sensitive to the share of costs passed on to rate payers, ranging from approximately US$0.9 trillion to US$2.8 trillion under pass-through assumptions of 95% to 85%, respectively, in the 1.5 °C scenario alone. By comparison, stranded assets increase by 25% to 60% when adopting lower and more geographically differentiated cost of capital estimates (Supplementary Table 5), as lower discount rates in many major economies raise the present value of foregone revenue (Supplementary Fig. 1). Although the absolute value of stranded assets increases under more ambitious climate scenarios and under lower-pass through or discount rates, the proportional concentration among power-generator operators (for example, fuel, region and ownership) of these assets remains relatively unchanged (Supplementary Fig. 1), indicating that these financing assumptions strongly affect magnitude but not the relative distribution of risks. Not shown in the figures, a well-known metric of market concentration, the Herfindahl–Hirschman index (HHI), indicates that ownership of stranded coal plants is substantially more concentrated (0.0142) than natural gas plants (0.0048; Methods).

Fig. 1: Value of fossil-fired electricity generators worldwide stranded under lower climate mitigation scenarios.
Fig. 1: Value of fossil-fired electricity generators worldwide stranded under lower climate mitigation scenarios.
Full size image

a–c, Stacked bars show the additional costs to fossil-fuel-fired electricity generators resulting from median regional carbon prices (n = 25 per region) under scenarios that limit the increase in global mean temperature to 1.5 °C, 2.0 °C, 2.2 °C and 2.6 °C. Bars represent median values, with error bars showing the range of stranded assets using minimum and maximum carbon prices within each region. Roughly 75% of the total added costs are related to coal-fired generators (a), more than half of the total are attributable to China (b) and more than a quarter are associated with power plants owned by just ten companies (c). d–f, The stranded-asset amounts under the same carbon pricing scenarios, assuming a 90% pass-through rate, with error bars showing the range of results calculated with pass-through rates of 85% and 95%. Coal power operators stand to lose around US$1.4 trillion over the operational lifespans of their plants (d). Similarly, China is projected to lose around US$1 trillion (e) and just ten companies could collectively lose half a trillion dollars (f) under a 1.5 °C scenario. ROW, rest of world; T, trillion.

Source data

Of the 100 companies that own the most stranded assets by value, the very largest are state-owned (denoted by an asterisk below) and the top 25 companies cumulatively hold upwards of US$770 billion in stranded assets and are responsible for emitting 4.0 GtCO2 per year—nearly 11% of global CO2 emissions in 2024. Among coal plants, these companies include China’s five largest power producers: Huaneng Group*, Datang Corporation*, Huadian Group*, China State Power Investment Corporation* and China Energy Investment*, which cumulatively account for US$79 to US$134 billion of coal assets stranded by the 2.0 °C scenario (Fig. 2a and Supplementary Fig. 2a). Beyond China, NTPC Limited* (owned by India’s Ministry of Power), Korea Electric Power* (South Korea), Perusahaan Listrik Negara* (Indonesia) and Adani Enterprises LTD (India) also hold around US$22 to US$33 billion in stranded coal-burning assets (Fig. 2a and Supplementary Fig. 2a). Stranded coal assets in China and India are large primarily because these countries are operating relatively new power plants: in both countries, the capacity-averaged age of coal-burning power plants is <15 years (Fig. 2b). The largest owners of stranded gas- and oil-burning plants include Egyptian Electric Holdings* (Egypt), Saudi Electricity* (Saudi Arabia), Nextera Energy (USA), Ministry of Electricity Iraq* (Iraq), Korea Electric Power (South Korea), Calpine (USA), CLP Group (China) and Engie S.A. (France; Fig. 2c). These companies operate large and/or young fleets of power plants (Fig. 2d).

Fig. 2: Top 100 companies by stranded-asset value in the 2.0 °C scenario.
Fig. 2: Top 100 companies by stranded-asset value in the 2.0 °C scenario.
Full size image

Circles show the value of stranded generation assets and related annual emissions for each of the 100 companies with the greatest total value of stranded assets, separated according to the type of fossil fuel burned by generators (rows). The size of the circles further indicates the fossil-electricity-generation capacity of each company and colours denote the country or region. a,c, Comparing the stranded assets of each company with the annual emissions from its power plants for coal (a) and gas (c). b,d, The average remaining operating life of those coal (b) and gas (d) plants until retirement.

Source data

The degree to which ownership of stranded fossil-fuel-burning power-plant assets is concentrated is shown in Fig. 3, which shows the share of fuel-specific emissions related to generators stranded under a carbon price consistent with a 2.0 °C scenario. Roughly 30% of CO2 emissions from coal plants are attributable to power plants owned by 10 companies (Fig. 3a,c) and more than half of CO2 emissions are from 50 companies. Ownership of stranded gas-burning power plants is more evenly distributed, with 100 companies owning gas-fired generators that emit half the emissions from natural gas- and oil-burning power plants (Fig. 3b,d). Some key assumptions underlying our engineering-based estimates of stranded assets include the operational capacity factor of individual fossil-fuel-burning power plants, carbon price trajectories, cost of capital estimates and pass-through rates. We test our results using lower and more granular cost of capital estimates (Supplementary Table 5) and observe an increase in stranded assets, but the relative patterns of stranding across firms, fuels and regions remain largely consistent (Fig. 3 and Supplementary Table 6). Figure 4 illustrates how estimated stranded-asset values vary under different power-plant operating conditions and various carbon pricing scenarios. We find that higher carbon prices (more stringent climate policies), extending fossil-fuel-burning power-plant lifespans and increased capacity factors all raise the value of stranded assets, but the relative importance of these variables is somewhat different for fossil-fuel-burning power plants burning different fossil fuels (Fig. 4). Coal-burning plants are the most sensitive to carbon price as they generally release more CO2 per unit of electricity generated and natural gas power plants are more evenly sensitive to both carbon prices and capacity factor (Fig. 4a,b). Similar patterns are observed with respect to power-plant operational lifespans, where carbon prices have a more pronounced impact on the stranded-asset values of coal power generators compared with gas-fired plants (Fig. 4d,c). Despite modelling assumptions, we observe that the distribution of stranded-asset risks among regions and companies remains largely consistent and companies with the greatest risk of stranded assets are those that have invested in and deployed the most fossil-fuel-burning power-plant capacity most recently.

Fig. 3: Corporate stranded assets by fuel and emissions.
Fig. 3: Corporate stranded assets by fuel and emissions.
Full size image

We organize stranded power-generating assets by value and show total power-sector emissions (x axis) by fuel (columns) and region (colours). The height of each bar represents the total stranded assets of each parent company under a 2 °C scenario, while the width of the bar indicates the proportion of the emissions of a company in relation to the total CO2 emissions from the power sector for a given fuel class. a,b, Stranded-asset values for coal (a) and gas (b) using a uniform global discount rate. c,d, Results derived using country-specific cost of capital estimates for coal (c) and gas (d). B, billion; K, thousand; M, million.

Source data

Fig. 4: Sensitivity of stranded assets to assumed carbon prices and power-plant operating conditions.
Fig. 4: Sensitivity of stranded assets to assumed carbon prices and power-plant operating conditions.
Full size image

Contours show how the value of stranded coal and gas power plants changes under various global carbon pricing scenarios which extend beyond the values used in lower warming scenarios. Carbon prices reflected on the y axis represent the maximum carbon price achieved under each case. a,b, The influence of the capacity factor (x axis) on stranded-asset values for coal (a) and gas (b). c,d, Illustrating how our estimates of stranded assets change under different power-plant operating lifespans (x axis) for coal (c) and gas (d).

Source data

Discussion

Uncertainties as to the timing and stringency of climate policies have encouraged continued investments in fossil-fuel-fired electricity generation, even as international climate goals (for example, limiting the increase of global mean temperature to around 2.0 °C) would entail the costly early retirement of a substantial portion of existing fossil infrastructure2,19,29,31. Here we develop an engineering-based estimate of stranded assets for the power sector and estimate the global total to be in the trillions of dollars over the next 40 years, more than half of which is owned by just 100 companies. Previous studies have estimated the economic impacts of stranded fossil fuel assets at national or global scales, reporting losses ranging from approximately US$500 billion to over US$4 trillion, depending on the stringency and timing of the implemented climate policy14,22,23,24,25,32. Our results are broadly consistent with these estimates, despite differences in methodological approach and the framing of stranded assets. Where previous studies assess stranded assets at the sector, country or financial system levels, our analysis identifies stranded assets at the company level and reveals that financial exposure is concentrated among a relatively few firms, particularly those with substantial investments in coal-fired power generation. While our estimates provide a benchmark for financial exposure under stringent climate policy, real-world asset stranding is likely to unfold in a more uneven environment, due to regional differences in policy, electricity market structures and financing conditions and thus may be more regionally differentiated than what is observed here. Still, our results may be useful for the growing number of investors and regulators interested in assessing policy-related transition risks18,33. In contrast to general statements about the potential costs of climate and energy policies, such as those made pursuant to the requirements of the Task Force on Climate-Related Financial Disclosures34, we have quantified the value of the assets of companies at risk of being stranded under different climate goals using publicly available information and transparent assumptions.

Under the most ambitious climate mitigation scenarios, we find stranded assets increase in companies whose power holdings have capacity-weighted average operational life of less than 20 years, most of which are owned and operated by China’s big five energy-generating companies—China Huaneng Group corporation*, China Datang Corporation*, China Huadian Corporation*, China Energy Investment* and China State Power Investment Corporation*. These state-owned enterprises are managed by the State Asset Supervision and Administration Commission (SASAC) and have been at the forefront of China’s energy and economic growth in recent decades. Cumulatively, SASAC manages 331 GW and 49 GW of coal- and gas-fired capacity, respectively (around 28.1% of China’s total electricity-generating capacity) and together, the power plants owned by these companies produce roughly 15% of the committed emissions of the global power sector. Consequently, China’s state-owned power infrastructure presents a unique opportunity and challenge for global decarbonization efforts as state-owned enterprises are largely incentivized by public service mandates rather than typical economic frameworks and are therefore more insulated from broader market forces35,36. Thus, the economic structure of state-owned enterprises can ultimately hinder or propel green technology investment relative to private firms, but is largely dependent on the regulatory directives of the state. Still, additional costs to state-owned power-generating enterprises constitute a public policy problem for national governments. Depending on the public objectives of those entities, costs may not be borne entirely by the enterprise itself but might be redistributed more generally across the population. While there is some evidence suggesting that state-owned enterprises may be more inclined than private firms to adopt environmentally friendly policies, the degree to which governments pursue policies counter to their economic objectives is relatively unknown and warrants further investigation35,37, particularly given that the scale of potential financial losses can also create strong incentives to resist more stringent climate policy. Given its notable role as a global producer of consumer goods38, the sudden financial collapse of the power infrastructure overseen by SASAC could have extensive financial implications beyond China’s borders and thus merits special consideration. By comparison, in Western economies, corporate power-generating entities and public utilities have largely lowered annual CO2 emissions by heavily investing in natural gas-fired power generation39. However, the substitution of aging coal-fired power generators with newer gas-fired power infrastructure has not substantially reduced cumulative emissions and has instead increased the transition risks for power-generating firms in the USA39 and European Union.

Limitations

While our findings offer valuable insights into the risks faced by specific companies, they are subject to several important caveats and limitations. First, our estimates of stranded assets are a result of financial conditions and policy trajectories that vary across countries and are likely to evolve over time, including climate policy (carbon pricing consistent with a specific warming target), cost of capital and the proportion of costs absorbed by individual companies. To the extent that the stringency of global climate policy is insufficient to meet lower warming targets, then our estimates will be correspondingly smaller. Similarly, our valuation of stranded assets is based on the assumption that fossil fuel power generators will continue to operate to the end of their working lives. While this assumption provides a useful upper bound for estimating stranded assets, it may overestimate operational persistence. In practice, there is considerable variation in power-plant lifespans due to regulatory design, maintenance requirements and economic conditions, which can lead to the early retirement of younger, more efficient generators or the extended operation of older plants40,41. Thus, to the extent that financial markets or asset owners already anticipate some tightening of climate policy over the lifetimes of assets or, conversely, reflect scepticism about the realization or stringency of climate policy, elements of this risk may already be priced into assets or appear on company balance sheets and thus be reflected in their current market valuation. Furthermore, substantial changes in fuel prices, evolving economic pressures or technological advancements may prompt earlier-than-expected retirements of fossil-fired-generating assets even in the absence of carbon pricing or climate-driven mandates. Therefore, given the complex and evolving factors driving stranded assets, our results should not be interpreted as forecasts for corporate revenues, investment returns, market capitalization or current stock valuation. Nor should our estimates be interpreted as unrealized capital that would automatically accrue to firms if more ambitious climate policy efforts were ultimately abandoned, as current company valuations already embed expectations about future profitability amid evolving macroeconomic conditions. Nonetheless, our results consistently show that the same companies hold the largest share of stranded assets across various climate mitigation policy trajectories and financing assumptions, indicating that these entities are most at risk of asset stranding.

It should also be noted that we assume companies will internalize some of the costs associated with carbon pricing, rather than transferring all additional expenses to their rate payers. In some cases, only a smaller portion of the carbon costs will be passed on to consumers, while in less competitive markets, all these costs can be passed through to rate payers42. The degree to which companies can pass costs on to consumers will depend on the development of carbon-free power generation, consumer demand response to increasing energy prices and regulatory processes that determine local electricity rates43,44. Our analysis attributes the added costs from carbon pricing as stranded-asset exposure to corporate owners, although we recognize that this does not fully capture the distribution of financial risk or liability. Utilities frequently use risk-sharing strategies, such as project-level non-recourse debt, leasing structures or regulatory cost of service frameworks, which shift financial exposure to lenders, lessors or consumers. However, these arrangements are often private and inconsistently disclosed, making them difficult to incorporate into asset-level modelling. Additionally, the timing at which carbon pricing is implemented also influences our results, as the financial risk of a company diminishes as its fossil-fired power fleet ages and increases as new carbon-emitting power plants come online. Consequently, the scale of stranded assets held by each corporate entity is likely to fluctuate over time. Although many countries, such as China and India, have nascent carbon pricing mechanisms, others, including the USA, lack such programmes at the national level.

Additionally, carbon policies that are more substantially geographically differentiated than the regional estimates used here, for instance, to reflect the differentiated responsibility of individual countries as opposed to economic regions, would produce different geographic distributions of stranded assets. Nonetheless, the carbon price schedules we adopt reflect cost-minimizing pathways for achieving more ambitious climate mitigation targets and the central pass-through rates used in this study are representative of those observed in more-established carbon markets within the electricity sector. If, instead, countries implement non-price instruments to achieve their emissions goals, the associated costs and potential for stranded assets will be correspondingly higher. Finally, our analysis is grounded in projections of power-plant revenues and costs based on historical electricity-generation returns. However, projected revenues and input costs are subject to change, with precise operational conditions, maintenance and fuel costs influenced by dynamic future conditions that are unknown and very difficult, if not impossible, to model. Yet changing costs and revenues will probably alter the operating schedules of future power generators.

Conclusions

Our results suggest that a substantial fraction of the existing fossil-fired generating assets of these companies could be stranded under current climate mitigation targets. However, while these costs pose a large financial risk to these companies, the monetary value of these stranded assets pales in comparison to the social and natural damages associated with unmitigated anthropogenic climate warming3,4,5,6,45,46,47. A sudden increase in power-generating costs from the implementation of carbon pricing can result in sizeable financial losses for companies that are heavily invested in carbon-intensive industries48. Thus, companies that persistently invest and deploy new CO2-emitting infrastructure without anticipating future tightening of climate policies face higher financial transition risks and ultimately increase the likelihood of unrecoverable previously anticipated projected returns (stranded assets). While gas-fired generators are likely to play a growing role in the clean energy transition because of their flexibility in supporting variable renewable energy, their profitability would still decline relative to a no-policy baseline as carbon prices rise. Therefore, to reduce stranded-asset exposure and align with climate mitigation trajectories, power-generating companies should strategically decommission or mothball their existing fossil-fuel-generating fleets.

Methods

We derived the corporate and state owners of 98% of individual operating fossil-electricity-generating plants from the global coal-plant tracker, global oil- and gas-plant tracker and global energy ownership tracker datasets obtained from the global energy monitor database. We omitted power plants from our calculations in the limited cases where the percentage of ownership was not listed within the data. We obtained technology-specific estimates of power-plant operating costs from the Fifth Assessment Report of the Intergovernmental Panel on Climate Change Working Group III (IPCC WGIII) and future carbon prices consistent with lower warming scenarios (1.5 °C, 2.0 °C, 2.2 °C and 2.6 °C scenarios) are obtained from the shared socioeconomic pathways database28,29,30. All monetary values are expressed in constant US$(2005), which is consistent with the carbon pricing estimates listed in the shared socioeconomic pathways database. While this approach ensures comparability across countries and time, it does not capture the potential impact of exchange rate fluctuations on historical or projected costs.

To quantify future profits, we project the expected generating income and costs of each power plant (i) individually over the anticipated operational life of the plant and assign profits to their respective parent company.

$$\mathrm{Generator}\,{\mathrm{profits}}_{{it}}={G}_{{it}}{\mathrm{EP}}_{{it}}-{G}_{{it}}{\mathrm{FC}}_{f}{\vartheta }_{i}-{{\rm{O}}\& {\rm{M}}}_{{it}}-{\mathrm{IC}}_{{it}}$$

In our study, generator profits are defined as the operating income earned from electricity generation over a period (t), calculated as power-plant revenue minus the cost of production. Income from production is determined by multiplying the annual electricity generation of each plant (G) with the regional price of electricity (EP) and subtracting the generating costs. Generating costs are calculated by multiplying annual electricity generation (G), global average fuel costs (FC) and a fuel-to-electricity conversion term (ϑ), in addition to the global operations and maintenance costs (O&M), as well as the investment costs (IC). Global average, technology-specific power-plant operational costs are obtained from the IPCC WGIII report49, while regionally distinct electricity prices are obtained from GCAM v.7 and although more granular data could improve accuracy in specific contexts, our approach allows for consistency and comparability across countries. In line with the IPCC approach, we include capital investment costs in our estimates. Although typically considered sunk and not expected to influence operational decisions, these costs reflect ongoing financial obligations that may persist for debt-financed infrastructure. We adopt a geographically uniform discount rate of 10%, consistent with the International Energy Agency (IEA) estimates for a high-risk market50,51,52, to reflect investor expectations under a riskier operating environment associated with climate policy and given the absence of a globally comprehensive cost of capital dataset. However, to test the sensitivity of our results against more heterogeneous financing conditions, we use lower regional- and country-level discount rates to reflect lower perceived investment risks and financing costs. Country-level cost of capital values are taken from ref. 53, which reports estimates for about 100 countries. These values are aggregated according to the regional classifications used in the carbon pricing data, with regional averages assigned to countries that have missing values. Additionally, as individual power plants age, the operational and maintenance costs typically increase such that these costs are not well covered by the generated revenues, resulting in their closure historically around 30 years for natural gas and 40 years12,54,55,56,57 for coal-fired power-generating infrastructure. Therefore, we allow for existing power infrastructure to retire in line with their historical norms, except for cases where a projected retirement date is explicitly listed in the power-plant data.

To quantify the impacts of climate mitigation policy on fossil-fired power-plant profits, we first calculate the net present value of these generating assets (i) on the basis of the operating income for each generator. We allow firms to retire these plants following historical conditions and do not explore the ramifications of the financial liabilities this may cause. Additionally, to avoid abrupt retirements of power-generating assets, we gradually raise the annual carbon price on an annual basis in line with the carbon pricing used by IPCC climate projection models, which use the shared socioeconomic framework30 (Supplementary Tables 1–4).

$${{\rm{NPV}}}_{i}=\displaystyle \mathop{\sum }\limits_{t=0}^{{\rm{LT}}}\frac{{\rm{Generator}}\,{\rm{profits}}}{{(1+r)}^{t}}$$

We translate future climate trajectories (1.5 °C through 2.6 °C) into stranded assets by incorporating regionally specific carbon price paths used by integrated assessment models for climate mitigation scenarios58. Carbon prices translate directly into increased input costs for fossil fuel plants, diminishing future profit projections. Therefore, we modify power-generator profits by including the additional accrued costs of electricity generation.

$$\begin{array}{l}{\mathrm{Modified}\,\mathrm{generator}\,\mathrm{profits}}_{it}\\ ={G}_{it}{\mathrm{EP}}_{it}-{G}_{it}{\mathrm{FC}}_{f}{\vartheta }_{i}-{G}_{it}{\mathrm{ECP}}_{t}(1-\mathrm{PTR})-{{\rm{O}}\& {\rm{M}}}_{it}-{\mathrm{IC}}_{it}\end{array}$$

Where E is the total annual emissions for each plant, PTR is the uniform pass-through rates and CP is the regionally based carbon pricing at the time (t), which achieves the climate emission scenario. Since most countries lack established carbon markets, there is limited empirical evidence available to determine national-level carbon price pass-through rates. Under conditions of perfect competition and full market efficiency, carbon costs would almost certainly be fully passed through to electricity prices, resulting in a 100% pass-through rate and no direct cost burden to fossil fuel generators59. In practice, however, observed pass-through rates vary across markets because of various factors such as market structure, regulatory interventions, the availability of alternative generation capacity and uncertainty in emission cost projections60,61. As a result, we adopt a range of 85% to 95% pass-through rates based on values reported in the existing literature42,62,63. Annual emissions are calculated for each individual unit from equations that vary slightly based on their fuel type. We obtain the emission equations from the Global Energy Monitor and IEA estimates.

$${{\rm{Annual}}\,{{\rm{CO}}}_{2}\,{\rm{emissions}}}_{{\rm{Coal}}}={G}_{{it}}\times {{\rm{HR}}}_{i}\times {{\rm{EF}}}_{i}\times C$$
$${{\rm{Annual}}\,{{\rm{CO}}}_{2}}\,{\rm{emissions}}_{{\rm{Gas}}}={G}_{{it}}\times {{\rm{LCEF}}}_{i}$$

Where HR is the technology-specific heat rate, EF denotes the emission factor of coal plants and LCEF represents the life-cycle emissions factor associated with natural-gas power plants. In our study, we define stranded assets as the difference in future asset income with and without the implementation of a climate policy that limits warming:

$${{\rm{Stranded}}\,{\rm{assets}}}_{i}=\displaystyle \mathop{\sum }\limits_{t=0}^{{\rm{LT}}}\frac{{{\rm{GP}}}_{i}-\widehat{{{\rm{GP}}}_{i}}}{{(1+r)}^{t}}=\displaystyle \mathop{\sum }\limits_{t=0}^{{\rm{LT}}}\frac{{G}_{{it}}{{\rm{ECP}}}_{t}(1-{\rm{PTR}})}{{(1+r)}^{t}}$$

We aggregate unit-level monetary income over the operational lifetime of the power generator and assign these losses to their parent company (corporate or state-enterprise owners) and select the top 100 companies by stranded-asset holdings. To investigate the national economic risk from stranded fossil-fired capacity, we quantify and assign the monetary value of these losses within the borders of each nation. We quantify the concentration of company-level stranded assets using the HHI, calculated as the sum of the squared shares of stranded assets relative to the total stranded assets for each fuel type. The HHI reflects the degree of concentration, with higher values indicating that a greater proportion of stranded assets is concentrated among fewer companies. Our analysis of stranded assets relies on publicly available data used by integrated assessment models for carbon prices28,29,30, which are commonly used in climate change forecasts and by the broader scientific community and historical average power-plant operating conditions for each individual unit. Nonetheless, we recognize that the value of stranded assets is determined by the carbon pricing amount and is also influenced by the use of power generators. Thus, we test the sensitivity of our results by varying the carbon price trajectory incrementally per year until the maximum price of US$0 to US$1,000 per ton is reached and fluctuating the capacity factor from 20% to 90%. We report the implications of these variations on global fossil-fired power-generator stranded assets in our sensitivity analyses. In a separate sensitivity test, we use historical capacity-factor estimates and instead vary the power-generator operating life and report these results. We subsequently repeat our sensitivity test, but instead of gradually increasing carbon prices over time, we apply the maximum carbon prices immediately and uniformly for the entire duration of the lifespans of the power infrastructures and show its effects on stranded-asset values. Overall, our approach enables us to assess the potential economic impacts of stranded fossil-fired power-generation assets, providing valuable insights into the financial risks faced by corporate entities and the challenges associated with transitioning to a low-carbon energy system.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.