Spatial mismatches constrain high-value utilization of retired batteries for decarbonization in China

Abstract

The rapid growth of electric vehicles is generating a large and expanding stream of retired batteries. High-value pathways, including power storage applications and direct recycling, deliver substantially greater economic returns and greenhouse-gas reductions than conventional metallurgical processing. Yet the spatial separation between regions where batteries retire and centers of high-value demand threatens to leave much of this potential unrealized. Here we show, using a province-level integrated model driven by vehicle- and company-level data, that spatial mismatches lock six billion tons of climate gains from high-value utilization of retired vehicle batteries in China between 2020 and 2050. An optimal strategy that prioritizes high-value pathways, expands collection and treatment infrastructure, and establishes efficient interprovincial transport networks unlocks these gains together with cost savings of nearly four thousand billion Chinese yuan relative to a business-as-usual baseline. These results demonstrate that early, coordinated planning can convert spatial barriers into scalable climate and economic benefits, offering a transferable pathway for other rapidly electrifying markets.

Graphical abstract

Highlights

• •Spatial mismatches lock six billion tons of climate gains from retired vehicle batteries.
• •Up to 87 percent of cost-saving potential remains locked without interprovincial transport.
• •High-value pathways still outperform metallurgical disposal by three to five times in cost.
• •Optimal planning unlocks the full six billion tons of carbon dioxide equivalent reductions.
• •The same strategy delivers nearly four thousand billion yuan in cumulative cost savings.

1. Introduction

Achieving global climate goals requires large-scale electrification across sectors, with electric vehicles (EVs) and battery energy storage systems (BESSs) serving as pivotal technologies [1,2]. New battery demand is projected to surge rapidly, intensifying concerns over supply constraints and environmental and climate externalities [[3], [4], [5]]. At the same time, the number of retired EV batteries is expected to grow dramatically [6,7], potentially causing pollution and safety hazards in the absence of effective end-of-life (EOL) management [8,9]. Addressing these challenges requires a circular-economy approach that integrates retired batteries into recycling and reuse pathways [10]. The reverse logistics system for retired batteries, which involves systematic collection, transport, and EOL processing, is emerging as a new sector with the potential to deliver synergistic environmental and economic benefits [11,12].

Many countries have enacted policies to support the growth of this emerging sector [13,14]. Global EOL battery processing capacity surpassed 2 million tons (Mt) in 2023 but remains almost entirely dependent on metallurgical recycling (MR) [15], which is favored for its established infrastructure and the widespread demand for recovered materials in other sectors. However, power storage applications and direct recycling offer higher-value EOL options, achieving greater economic returns and substantially higher greenhouse gas (GHG) mitigation potential than MR [16,17]. Ma et al. reported that high-value options can improve profits by 19–58%, and cut emissions by 18% compared with MR approaches [16]. Nevertheless, the market for high-value utilization remains underdeveloped, primarily due to limited current demand for BESSs and for directly recycled EV batteries [18,19]. Logistical barriers to battery transport, along with the upfront investment required for high-value utilization infrastructure, further lead companies to favor established MR approaches [20,21]. Despite policy efforts to encourage this transition, progress remains slow (barriers to direct recycling and second-life reuse are detailed in Supplementary Texts S1–S2). As electrification is expected to further increase demand for new batteries, early reliance on MR may lock companies into inflexible disposal pathways, hindering the development of business models built around high-value utilization and limiting their ability to realize long-term environmental and economic benefits.

Existing studies at multiple scales show that high-value utilization achieves greater carbon reduction and cost savings than MR [[22], [23], [24]]. For example, Ciez et al. reported that direct recycling can reduce emissions and remain economically competitive at the process scale [14]. Tao et al. conducted a life-cycle assessment and found that power storage applications can reduce the carbon footprint by 8–17% [23]. Jiang et al. evaluated the economic and GHG mitigation benefits of transitioning to high-value utilization at the country scale [25]. Together, these studies confirm the substantial environmental and economic advantages of high-value utilization and underscore the urgent need to advance this emerging sector. However, a key constraint on high-value utilization is the persistent spatial mismatch between battery-retirement locations and demand centers. Retired battery supply tends to be concentrated in economically developed, densely populated regions with high EV ownership, whereas demand is concentrated in regions with battery production capacity or high renewable energy potential [17]. This spatial mismatch leads to numerous batteries being stranded or wasted in retired regions, thereby limiting their effective utilization in demand regions [[26], [27], [28]]. For example, approximately 465,000 tons of retired batteries were wasted locally in China in 2022 [29], turning valuable resources into waste. As electrification expands, spatial mismatch is likely to intensify, potentially limiting the realization of projected GHG mitigation and cost benefits highlighted in the literature. The impact of spatial supply–demand mismatches on emerging high-value disposal markets has received insufficient attention. However, a few preliminary studies have explored this issue [21,30]. By mapping the future spatial distribution of retired EV batteries and BESS demands, Jiang et al. [21] confirmed the existence of supply–demand mismatches in China. Tian et al. [30] investigated how to manage existing disposal capacity, which is predominantly MR, to maximize economic and environmental benefits. Nevertheless, as China's battery disposal sector is poised for a significant surge in both demand for high-value utilization and construction of corresponding facilities, the long-term strategy for this sector under this new paradigm of rapid expansion and spatial mismatch remains inadequately studied. How this transition is constrained by supply–demand spatial mismatch, the associated economic and environmental impacts, and potential optimization pathways remain underexplored.

To address this gap, we examine the GHG mitigation and cost-saving potential of shifting retired battery disposal from MR to high-value utilization pathways. As the world's largest EV market and an emerging leader in battery recycling and reuse [31], China offers a critical case for understanding the development and challenges of the emerging high-value utilization market. We develop a province-level coupling model for China's retired battery disposal sector that captures the spatial distribution of retired batteries and demand centers and quantifies the sector's life-cycle GHG emissions and costs. Our findings show how spatial mismatches between supply and demand can be addressed to unlock the potential of high-value battery utilization as the market matures. These insights are also relevant to other rapidly electrifying markets, such as the United States [26] and Europe [27], which face comparable opportunities and challenges.

2. Materials and methods

2.1. Integrated model framework

We developed a province-level integrated model to project the supply, demand, and disposal capacities of China's retired batteries from 2020 to 2050 (Fig. 1). The integrated model comprises three main modules: supply, demand, and disposal. The supply module outputs each province's capacity for retired battery supply; collected batteries enter the disposal system, whereas uncollected batteries are treated as waste. The demand module outputs each province's demand capacity for retired batteries. Based on these supply and demand outputs, the disposal system quantifies the degree of spatial mismatch and develops interprovincial transportation plans. We then designed future development scenarios for China's retired battery disposal sector to assess the impacts of battery collection rates and EOL pathway choices, and finally assessed GHG emissions and costs using a life-cycle technoeconomic assessment. Each component is described in detail below.

Fig. 1.

Integrated modeling framework for the supply, demand, and end-of-life management of retired electric-vehicle batteries. The supply module uses a dynamic fleet-stock model to project provincial retired-battery supply from 2020 to 2050. The demand module estimates demand for retired batteries in power-storage applications based on projected solar and wind power generation, as well as demand for recycled batteries associated with new-car battery production. The disposal-system module distinguishes collected from uncollected batteries, matches provincial supply with demand, optimizes interprovincial transport, and allocates collected batteries among metallurgical recycling, power storage application, and direct recycling. The analytical framework evaluates alternative collection rates and end-of-life pathway choices, quantifies spatial mismatch using the Gini coefficient, optimizes transport, and assesses life-cycle greenhouse-gas emissions and system costs. A Monte Carlo approach is used to identify the optimal strategy. Solid arrows indicate flows of collected batteries through the formal management system; dashed arrows indicate uncollected batteries or waste flows.

2.2. Retired battery supply module

Retired batteries are lithium-ion traction batteries withdrawn from vehicle use after their capacity declines to approximately 80% of their initial value [17]. We developed a dynamic fleet stock model based on material flow analysis—a method well-established and widely applied [[32], [33], [34]]—to simulate China's EV fleet from 2020 to 2050. Based on this model, we derived annual sales of battery EVs and plug-in hybrid EVs for each Chinese province. We further categorized these EVs by body type into sedans, sport-utility vehicles, multipurpose vehicles, and crossovers. We quantified the representative capacities of nickel manganese cobalt and lithium iron phosphate batteries used across approximately 600 EV models currently sold in China, mapped to these eight categories. This data-driven approach, linking EVs to battery chemistry attributes, follows the method validated by Dong et al. [6]. EV sales and assembly data are sourced from the Marklines database (https://www.marklines.com/cn/). Retired batteries arise from two sources: batteries that have reached EOL capacity (approximately 80%) and batteries retired alongside EV scrappage. For the first source, remaining capacity at retirement was estimated using the lifespan-curve approach [6]; for the second, retirement volumes were derived from our dynamic fleet stock model. Considering the significant impact of temperature on EV lifespan, we adjusted the battery lifespan function for each province to monthly mean temperatures using the Arrhenius equation, parameterized with the temperature-lifetime curve reported by Lam et al. [35]. We additionally adjusted battery lifetime according to each province's average annual driving distance to capture the impact of vehicle usage intensity [36]. Detailed modeling procedures and parameter settings for the supply module are provided in Supplementary Text S3.

2.3. Retired battery demand module

Retired battery demand arises from three EOL pathways: MR, power storage application, and direct recycling. Given the widespread distribution of MR facilities across China [21] and their capacity to serve multiple sectors, we assume that each province's MR demand and treatment capacity exceed its maximum supply of retired batteries. We considered only hydrometallurgy for MR, as pyrometallurgy cannot efficiently recover lithium [6]. We selected energy storage as the representative path due to its economic viability and market scalability [37]. China's Interim Measures for the Management of Recycling and Comprehensive Utilization of Retired New Energy Vehicle Power Batteries (https://www.miit.gov.cn/gyhxxhb/jgsj/cyzcyfgs/bmgz/jdcjxl/art/2026/art_392462fdc40c415ea4a4129cac3028c2.html) explicitly prohibits the use of retired power batteries in electric bicycles, whether directly or after processing. We excluded communication base station backup power, as it is a mature, standardized application with limited incremental deployment potential [38]. Only lithium iron phosphate batteries are considered for power storage applications; nickel manganese cobalt batteries are excluded owing to safety and durability concerns [39]. Provincial wind and solar generation under China's 2060 carbon neutrality pathway is drawn from our previous Global Change Assessment Model (GCAM)-based study [40]. The required storage battery stock is estimated using a pairing-coefficient method for wind and solar power, following Geng et al. [39]. We then estimated the demand for new and retired batteries using battery lifetime curves and the material flow analysis method. We neglected temperature effects on the lifetime of storage batteries, as thermal management systems are standard in storage applications [41]. The total demand for vehicle batteries comprises both initial installations in new vehicles and replacement batteries for aging fleets. China enforces an extended producer responsibility policy, mandating battery manufacturers to recycle retired batteries [42]. We projected China's total vehicle battery demand using our dynamic fleet-stock model and allocated it to provinces based on current battery production capacity. Provincial production capacity was derived from data for 1021 Chinese battery manufacturers from the Parts and Components Suppliers Database in Marklines (https://www.marklines.com/cn/supplier_db), supplemented with business-operation information from TianYanCha (https://www.tianyancha.com/). Detailed modeling procedures and parameter settings of the retired battery demand module are thoroughly described in Supplementary Text S4.

2.4. Retired battery disposal system module

We proposed a life-cycle assessment framework for China's retired battery disposal system, comprising collection, inspection and dismantling, interprovincial transport, MR, power storage application, direct recycling, and waste stages (Fig. 1). Batteries retire from power storage application once their capacity drops from 80% to 40–50%, after which they enter MR [6,39]. Batteries produced through direct recycling have the same lifetime as new batteries, and direct recycling is permitted only once per battery over the entire life cycle. After EOL processing, components with no remaining value are sent to the waste treatment process. We proposed a transportation optimization model that minimizes transportation costs while clearing all surplus retired batteries. Interprovincial transport distances are defined as highway distances between provincial capital cities; intraprovincial collection transport distances are based on each province's average internal transport distance. We assumed transport distances for other processes were negligible and that each destination region would develop sufficient disposal capacity to meet demand for retired-battery treatment. Detailed modeling procedures and parameter settings for the disposal system module are provided in Supplementary Text S5.

2.5. Scenario configurations

We configured a business-as-usual (BAU) scenario and 12 alternative scenarios for the evolution of China's battery disposal sector from 2020 to 2050 (Table 1). In the BAU scenario, both collection rate and EOL pathway distribution remain at 2020 levels through 2050. In the alternative scenarios, collection rate increases at three levels: low, middle, and high. For each collection level, we further defined four scenarios for EOL pathway allocation: three scenarios prioritizing MR, direct recycling (DR), and power storage application (SU), respectively, and a balanced (BA) scenario promoting parallel development of DR and SU. Within each scenario, we adopt uniform national assumptions for collection rate and EOL pathway distribution to isolate differences in provincial implementation effects.

Table 1.

Scenario configurations for China's retired battery disposal sector.

Scenario group	Collection-rate trajectory	End-of-life pathway shares in 2050
Business-as-usual (BAU)	Stay at 25% 2020–2050	100% MR throughout 2020–2050
Low-MR	Low-speed growth to 48% by 2050	86% MR, 7% SU, 7% DR in 2050
Low-DR	Low-speed growth to 48% by 2050	37% MR, 6% SU, 57% DR in 2050
Low-SU	Low-speed growth to 48% by 2050	37% MR, 57% SU, 6% DR in 2050
Low-BA	Low-speed growth to 48% by 2050	37% MR, 32% SU, 32% DR in 2050
Middle-MR	Middle-speed growth to 72% by 2050	86% MR, 7% SU, 7% DR in 2050
Middle-DR	Middle-speed growth to 72% by 2050	37% MR, 6% SU, 57% DR in 2050
Middle-SU	Middle-speed growth to 72% by 2050	37% MR, 57% SU, 6% DR in 2050
Middle-BA	Middle-speed growth to 72% by 2050	37% MR, 32% SU, 32% DR in 2050
High-MR	High-speed growth to 95% by 2050	86% MR, 7% SU, 7% DR in 2050
High-DR	High-speed growth to 95% by 2050	37% MR, 6% SU, 57% DR in 2050
High-SU	High-speed growth to 95% by 2050	37% MR, 57% SU, 6% DR in 2050
High-BA	High-speed growth to 95% by 2050	37% MR, 32% SU, 32% DR in 2050

MR, metallurgical recycling; DR, direct recycling; SU, power storage application; BA, balanced scenario. Scenario values were adapted from Dong et al. [6]. The pathway shares apply only to batteries entering their first end-of-life pathway and exclude batteries that have previously undergone power storage application or direct recycling. Percentages may not sum to 100% because of rounding.

2.6. Life-cycle technoeconomic assessment and Monte Carlo approach

We integrated technoeconomic analysis with life-cycle assessment to systematically examine the GHG mitigation potential and costs of China's emerging retired battery disposal sector, incorporating GHG emissions from material inputs, electricity consumption, and infrastructure construction. Provincial electricity GHG emission factors are derived from our GCAM model, and cradle-to-grave GHG emission factors for new vehicle and energy storage batteries are sourced from Ma et al. [16]. For costs, we considered the standardized fixed and operational costs based on battery processing capacity for inspection and dismantling, power storage applications, and DR processes. Other processes, such as collection, interprovincial transport, and metallurgy, account only for standardized operational costs, as they rely on existing logistics systems (e.g., trucking for collection and interprovincial transport) and established MR capacity in China. Each province's existing MR capacity is generally sufficient to meet its own projected treatment demand (Supplementary Text S6). We accounted for practical constraints by incorporating additional compliance-related costs into the transport module, reflecting the use of existing road transport systems under realistic operating conditions without requiring additional capital investment. Because retired batteries cannot be stacked in transit owing to fire and shock risk, effective cargo space utilization is reduced to roughly half that of ordinary freight, incurring additional compliant transport costs of 100–300 Chinese yuan (CNY) ton⁻¹ (set at CNY 200 ton⁻¹ in this study) [43]. We estimated the environmental cost of uncollected retired batteries by treating them as untreated hazardous waste. Under China's Law on the Prevention and Control of Environmental Pollution by Solid Waste (https://www.mee.gov.cn/ywgz/fgbz/fl/202004/t20200430_777580.shtml), unauthorized dumping of hazardous waste is subject to a fine equivalent to three to five times the required disposal cost; we conservatively applied the lower bound of three times the disposal cost. All costs were discounted at an annual rate of 2%.

We used the Monte Carlo approach to explore province-specific combinations of collection rates and EOL pathways across 150,000 model simulations. This near-exhaustive approach enables a comprehensive understanding of the potential development pathways for China's retired battery disposal sector. We then identified the optimal solution based on the cost savings and GHG emission reductions across all simulations. Detailed modeling procedures and parameter settings of cost–benefit analysis and the Monte Carlo approach are provided in Supplementary Text S3.

3. Results

3.1. Cost-saving and GHG mitigation potential of high-value utilization approaches

We evaluated relative cost-saving changes in each alternative scenario against the BAU baseline (Fig. 2a). Scenarios prioritizing high-value utilization are projected to yield greater cost savings than MR-dominated scenarios. By 2050, MR scenarios are projected to achieve cost savings of up to CNY 49 billion, even at higher collection rates, whereas high-value utilization scenarios (DR, SU, and BA) are projected to achieve CNY 94–258 billion even at low collection levels. Thus, adopting high-value utilization strategies is projected to be more effective than simply increasing collection rates. Over 2020–2050, cumulative cost savings in high-value utilization scenarios are projected to be 2.9–5.4 times higher than in MR scenarios. In 2050, the maximum cost-saving gaps between MR-dominated scenarios and high-value utilization scenarios are projected to be CNY 226 billion, CNY 234 billion, and CNY 254 billion under low, middle, and high collection rates, respectively.

Fig. 2.

Cost-saving and GHG mitigation potential of scenarios from 2020 to 2050. Changes in cost (a) and GHG emission (b) of alternative scenarios compared to the business-as-usual (BAU) scenario. The three columns, from left to right, correspond to low, middle, and high collection rates, respectively. Coloured lines indicate the dominant end-of-life pathway: metallurgical recycling (MR; orange), direct recycling (DR; blue), power-storage applications (SU; pink) and a balanced allocation among pathways (BA; green). Negative values indicate cost savings or avoided GHG emissions relative to BAU. GHG, greenhouse gas; CNY, Chinese yuan; CO₂eq, carbon dioxide equivalent.

We then quantified changes in GHG emissions relative to the BAU scenario (Fig. 2b). Over 2020–2050, cumulative reductions under high-value utilization scenarios are projected to be 1.2–3.0 times those under MR scenarios. In MR scenarios, the carbon mitigation opportunity lost at low, middle, and high collection rates is estimated at 204, 303, and 401 Mt CO₂eq, respectively, relative to high-value utilization scenarios in 2050.

Overall, prioritizing high-value utilization over MR disposal in China could increase cost savings by approximately 3–5-fold and GHG emission reduction by 1–3-fold. These results highlight an urgent need for sectoral transition and strategic investments in high-value utilization pathways.

3.2. Supply–demand spatial mismatch for high-value utilization approaches

We quantified China's supply–demand gaps for power storage applications from 2020 to 2050 (Fig. 3a). In all scenarios prioritizing power storage application as the main EOL pathway, retired battery supply is projected to be sufficient to meet China's future energy storage demand; even at the low collection rate, available retired batteries exceed total demand. In the high-SU scenario, available surpluses are projected to reach 210 GWh by 2035 and 1755 GWh by 2050. However, shortages are projected to be concentrated in renewable-rich western provinces (e.g., Nei Mongol, Xizang, Qinghai, Xinjiang), whereas surpluses are projected to be broadly distributed across economically developed eastern provinces. By 2050, the shortage gap in Nei Mongol is projected to reach approximately 340 GWh, while Guangdong and Shandong each have surpluses exceeding 220 GWh.

Fig. 3.

Temporal and spatial mismatches between retired-battery supply and demand for high-value utilization. a, Battery capacity gaps associated with power storage applications. The left panel shows the national supply–demand gap from 2020 to 2050 across collection levels and end-of-life pathways; the right panels show provincial capacity gaps under the High-SU scenario in 2035 and 2050. b, Share of recycled batteries in demand and the corresponding battery-capacity gaps for direct recycling. The left panel shows the national share from 2020 to 2050, and the right panels show provincial capacity gaps under the High-DR scenario in 2035 and 2050. In the time-series panels, dotted, dashed, and solid lines represent low, middle, and high collection rates, respectively; colors denote the MR, DR, SU, and BA pathways, and the grey line denotes BAU. In the provincial panels, positive and negative values indicate battery surpluses and deficits, respectively. Insets show Lorenz curves, with the dashed diagonal indicating perfect spatial equality; G denotes the corresponding Gini coefficient. MR, metallurgical recycling; DR, direct recycling; SU, power-storage applications; BA, balanced allocation; BAU, business as usual.

We then assessed China's retired battery supply–demand gaps for DR from 2020 to 2050 (Fig. 3b). In all scenarios, retired batteries are projected to meet only part of recycling demand. Under the high-DR scenario, retired-battery supply is projected to meet only approximately 25% of the recycling demand by 2050, representing the most optimistic case. Despite this substantial national-level shortage, many provinces are expected to have surpluses: by 2050, those in Liaoning and Shandong provinces are expected to exceed approximately 150 GWh. Conversely, major battery-producing provinces are projected to face significant shortages, with gaps reaching 3800 GWh in Guangdong and 4800 GWh in Fujian.

We calculated the Gini coefficients for the high-SU and high-DR scenarios in 2035 and 2050 to evaluate the severity of supply–demand spatial mismatch (Fig. 3). A higher Gini coefficient value indicates more pronounced spatial mismatch across provinces [44]. In the high-SU scenario (Fig. 3a), the Gini coefficient is projected to rise from 0.6 in 2035 to 0.7 in 2050, reflecting growing spatial mismatch driven by ongoing electrification. The rising renewable generation in western provinces is projected to increase local battery storage demand, whereas rapid EV adoption is projected to produce a larger retired battery stock in eastern regions, widening the east–west gap. Under the high-DR scenario (Fig. 3b), the Gini coefficient is projected to decrease from 0.9 in 2035 to 0.8 in 2050, indicating a modest reduction in spatial mismatch. Hebei, Shandong, and Henan provinces are projected to shift from retired-battery shortages in 2030 to surpluses in 2050, driven by electrification growth, which is projected to increase retired-battery availability in production regions.

These results indicate that China's electrification progress is projected to intensify the retired battery supply–demand spatial mismatch for power storage applications, while slightly mitigating the mismatch in DR.

3.3. GHG mitigation and cost-saving potential losses due to spatial mismatch

We quantified the constrained cost-saving potential due to spatial mismatch across scenarios from 2020 to 2050 (Fig. 4a). In scenarios that prioritize power storage applications or direct recycling as primary EOL pathways (DR, SU, or BA), spatial mismatch is projected to constrain 50–87% of the cost-saving potential. When a power storage application is adopted as the primary EOL path, spatial mismatch is projected to be most severe, leaving 71–87% of cumulative cost-saving potential unrealized.

Fig. 4.

Losses in cumulative cost-saving and GHG-mitigation potential caused by spatial supply–demand mismatches, 2020–2050. a–b, Realizable cost-saving potential (a) and GHG-mitigation potential (b) under spatial constraints, expressed as percentages of the corresponding unconstrained potentials. c–d, Contributions of individual processes to cumulative cost changes (c) and GHG-emission changes (d) under the High-SU scenario, relative to BAU. e–f, Corresponding cumulative cost changes (e) and GHG-emission changes (f) under the High-DR scenario. In panels c–f, the grey dashed line separates outcomes without interprovincial transport (left) from the additional changes resulting from transport (right); red annotations report the net changes for each case. Negative values indicate cost savings or avoided GHG emissions relative to BAU. MR, metallurgical recycling; DR, direct recycling; SU, power-storage applications; BA, balanced allocation; BAU, business as usual; GHG, greenhouse gas; CNY, Chinese yuan; CO₂eq, carbon dioxide equivalent.

We further decomposed unrealized cost savings under the high-SU scenario (Fig. 4c). Local self-treatment of retired batteries is projected to save CNY 293 billion compared with the BAU scenario. These savings arose from reductions in the environmental costs of uncollected batteries and in the costs associated with the three EOL paths, whereas battery collection and the inspection and dismantling processes add costs. Interprovincial redistribution could unlock a further CNY 2003 billion in savings over 2020–2050, while incurring only CNY 48 billion in additional transport costs.

Scenarios prioritizing DR as the primary EOL path are projected to be less constrained in cost savings than power-storage-dominated scenarios, but considerably more constrained than MR scenarios, over 2020–2050 (Fig. 4a). Under the high-DR scenario (Fig. 4e), local self-treatment of retired batteries is projected to save CNY 666 billion compared with the BAU scenario, larger than that of the high-SU scenario. Interprovincial transport could unlock the remaining CNY 1943 billion in savings at an additional transport cost of CNY 59 billion, accounting for only 2.3% of the overall cost-saving potential.

We further quantified constrained GHG mitigation potential due to spatial mismatch across scenarios from 2020 to 2050 (Fig. 4b). In scenarios where DR, power storage application, or a balanced combination is the primary EOL path, 26–53% of GHG mitigation potential is projected to be constrained. For the high-SU scenario (Fig. 4d), local self-treatment mitigated 1593 Mt CO₂eq, with 1078 Mt CO₂eq unlocked by interprovincial transport at a cost of only 0.7 Mt CO₂eq in transport-related emissions, accounting for only 0.03% of the overall mitigation potential. Compared with the high-SU scenario, the high-DR scenario is projected to offer greater GHG mitigation potential (Fig. 4f). Local self-treatment of retired batteries is projected to enable 3756 Mt CO₂eq in GHG mitigation. The 1904 Mt CO₂eq of potential is projected to remain constrained due to spatial mismatch. Interprovincial transport is projected to emit only 0.9 Mt CO₂eq GHG, equivalent to just 0.01% of the overall mitigation potential.

These findings underscore the urgent need for integrated planning of collection systems, EOL path selection, and interprovincial transport to unlock constrained mitigation and cost-saving potential.

3.4. Optimal solution for China

We identified an optimal solution for China's retired battery sector to maximize GHG mitigation and cost-saving potentials from 2020 to 2050 (Fig. 5a). Relative to the BAU scenario, the optimal solution is projected to yield CNY 3988 billion in cost savings and 6161 Mt CO₂eq in GHG mitigation, outperforming all scenarios with uniform national rates on both metrics (Fig. 5b and c).

Fig. 5.

Optimal end-of-life management strategy and its cumulative cost and GHG outcomes from 2020 to 2050. a, Cumulative GHG-emission and cost changes relative to BAU for strategies sampled using Monte Carlo simulation. Grey points represent simulated strategies; symbol shapes denote low, middle, and high collection levels, and symbol colors denote the MR, DR, SU, and BA end-of-life pathways. The red rectangle marks the identified optimal strategy (minimizing cost changes). b–c, Contributions of individual processes to cumulative GHG-emission changes (b) and cost changes (c) under the predefined scenarios and optimal strategy. Bar colors identify the processes shown in the key, and open triangles indicate net changes. Negative values denote avoided GHG emissions in panel b and cost savings in panel c, relative to BAU. MR, metallurgical recycling; DR, direct recycling; SU, power-storage applications; BA, balanced allocation; BAU, business as usual; EOL, end of life; GHG, greenhouse gas; CNY, Chinese yuan; CO₂eq, carbon dioxide equivalent.

The optimal solution is projected to prioritize power storage application and DR as the primary EOL paths, which accounted for approximately 24% and 56%, respectively, of total collected batteries in 2050 (Supplementary Fig. S1a). Moreover, approximately 14% are projected to require MR processing for disposal in 2050. Despite extensive interprovincial transport, supply–demand gaps persist. For the power storage application path, Nei Mongol, Xinjiang, and Xizang are projected to face supply shortages in 2035, but only Xinjiang is projected to remain with a gap of approximately 34 GWh by 2050 (Supplementary Fig. S1b). Henan, Hebei, and Guangdong are projected to emerge as the primary exporters of retired batteries in 2050. For the DR path, 12 provinces are projected to face shortages in 2035, narrowing to Guangdong, Fujian, and Guizhou by 2050 (Supplementary Fig. S1c), with Henan, Shanxi, and Xinjiang as the main exporters.

4. Discussion

Retired battery disposal is an emerging sector with significant market potential. We find that China's high-value utilization of retired batteries can achieve cumulative cost savings and carbon reductions 3–5 and 1–3 times greater, respectively, than continued reliance on MR disposal. However, high-value utilization remains in its early stages, constrained by a supply–demand spatial mismatch that is likely to intensify as electrification deepens. Over 2020–2050, these spatial constraints are projected to leave 26–53% of the potential GHG mitigation and 50–87% of the potential cost savings unrealized. Planned interprovincial transportation can effectively relieve these constraints at minimal additional cost and emissions. Relative to the BAU scenario, our optimal solution is projected to deliver cumulative GHG mitigation of 6161 Mt CO₂eq and cost savings of CNY 3988 billion. These results highlight the importance of establishing clear development pathways early in the sector's growth to guide investment in infrastructure and transport networks and to avoid technological lock-in around prevailing MR preferences.

Previous national-level assessments of the GHG mitigation and economic benefits of high-value utilization [6,16] have relied on aggregate estimation, neglecting the spatial mismatch between battery-retirement sources and downstream demand. This methodological simplification overlooks critical real-world constraints on the performance of high-value utilization. However, large quantities of retired EV batteries remain underutilized in China due to a lack of nearby recycling and EOL facilities, resulting in resource waste and environmental pollution [45,46]. Our findings reveal that these spatial mismatches substantially undermine the theoretical benefits projected by previous literature, highlighting the need to integrate spatial planning into system design— a dimension that remains underexplored. By quantifying the extent of these limitations and demonstrating the effectiveness of targeted transport strategies, our study offers a more realistic and actionable basis for optimizing performance in practice.

Our optimal solution positions power storage application and DR as the primary EOL pathways, gradually replacing MR after 2030. Although this transition may reduce the short-term volume of batteries entering MR processing, it is unlikely to cause long-term shortages of critical battery metals. Batteries deployed in power storage applications typically retire within approximately six years and can subsequently be redirected into MR [47]. Similarly, regenerated batteries can also undergo MR processing at the end of their service life. Establishing a robust battery traceability system that clarifies EOL ownership and responsibility will be essential to ensuring both material security and the efficient circular use of battery resources [48].

However, DR faces both technical and economic barriers. Technical challenges include achieving nondestructive separation during pretreatment, minimizing residual impurities that degrade regenerated material quality, and accurately screening heterogeneous EOL batteries. Economically, current operating costs remain high, and market acceptance of regenerated materials is uncertain. Future progress hinges on low-cost, high-efficiency smart screening, uniform regeneration across varied degraded materials, performance-enhancing doping or coating, and supportive market policy [49].

We suggest establishing a streamlined national interprovincial transport network to enhance transport efficiency and minimize spatial supply–demand mismatch. Currently, the circulation of retired batteries in China faces significant barriers, particularly complex approval procedures for interprovincial transport. Because retired power batteries are classified as hazardous materials, their transfer between provinces requires joint approval from the ecological and environmental authorities of both origin and destination provinces. Transport also presents safety risks, including leakage, deformation, fire, and water damage [50]. Although specialized battery transport vehicles can mitigate these risks, such vehicles remain scarce in China. Requirements for certified personnel and unstable transport volumes further raise operational barriers, leaving enterprises with little incentive to engage in compliant transport. A stable, standardized transport network would both regulate market order among transport operators and improve safety and overall logistics efficiency. In practice, time delays, queuing effects, and inventory management may also affect the timing of battery flows, underscoring the need for coordinated policy support, improved operational planning by disposal facilities, and more efficient logistics management by transport operators.

Uneven development of collection networks and high-value disposal infrastructure across China is another critical challenge. Many provinces lack sufficient treatment capacity [51,52], often prompting batteries to be prematurely shredded to facilitate long-distance transport [53] and thereby precluding higher-value utilization. Moreover, with electrification-driven growth in demand for energy storage, especially in western regions with rapid deployment of renewable energy, there is an urgent need to develop localized infrastructure to support power storage applications. In these regions, encouraging adoption of EV models equipped with lithium iron phosphate batteries can facilitate the development of local power storage application systems. Addressing both transport and infrastructure challenges in a coordinated manner is essential to unlocking the sector's full potential and supporting the broader energy transition.

Market conditions also shape the effectiveness of retired-battery management. In China, owners may return retired power batteries to automakers through trade-in schemes or dispose of them independently [54]. However, limitations in transportation and EOL processing capacity drive many batteries into illegal channels, leading to inefficiency and price distortion [55]. Furthermore, the current disposal model for EVs still follows the traditional model for internal combustion engine vehicles, offering insufficient incentives for consumers to properly dispose of their vehicles [56]. China should therefore implement disposal subsidies for retired batteries and establish a tailored EV disposal model. Given the potential cost benefits of transitioning to high-value reuse pathways, relevant stakeholders should take responsibility for facilitating this shift. Market supervision authorities should also strengthen penalties for unqualified and unlicensed small-scale recycling operations, curbing illegal practices and fostering a more regulated, sustainable recycling market that benefits both legitimate disposal enterprises and car owners, and supports the overall scale-up of the disposal market. Alternative business models, such as battery banks and battery-swapping systems [57], remain in early development but may emerge as viable future solutions.

Although tailored retired-battery disposal can maximize GHG mitigation and cost-saving potential, its implementation in practice faces several challenges. These include heterogeneity in battery chemistries and designs [58], limited standardization for inspection and dismantling processes [59], and safety risks in power storage applications and recycling batteries [60]. Emerging battery technologies, such as solid-state batteries, may further reshape this sector, but require validation through real-world pilot programs and adaptive policy support [61].

Our uncertainty analyses (Supplementary Text S7) show that the main results are robust to uncertainties in key model parameters, including the discount rate, the environmental cost of uncollected retired batteries, the capacity utilization of retired battery treatment facilities, inter-province transport distance, and the energy mix. Nonetheless, the modeling framework could be strengthened by incorporating more comprehensive data, such as detailed information on disposal infrastructure, on-site measured cost parameters and emission factors, and the behavioral preferences of consumers and enterprises. For example, owing to a lack of operational data, we assume full capacity utilization across disposal facilities, whereas actual utilization is context-dependent. Policymakers and industry stakeholders should therefore adopt locally tailored strategies when planning disposal infrastructure and managing battery flows.

Our model also omits interprovincial MR processing, which is primarily driven by price factors [62] that lie beyond this study's scope. In addition, we considered road transport as the dominant mode given current market conditions. Multimodal options such as rail and waterways may further reduce transport costs and emissions [63,64], suggesting that our estimates are conservative and highlighting the need to improve intermodal coordination and logistics planning for the transportation of retired batteries. In practice, disposal capacity may not immediately align with spatial demand, and its expansion requires time and coordinated investment, potentially constraining short-term battery reallocation and delaying realization of the benefits identified here. Despite these limitations, our findings underscore the importance of shifting retired battery disposal from MR to high-value utilization, offering insights relevant to the sustainable use of retired battery resources and the broader electrification transitions worldwide.

5. Conclusion

High-value pathways for retired vehicle batteries have the potential to deliver larger economic and climate benefits than metallurgical recycling alone. Our analysis, however, indicates that spatial mismatches between retired-battery supply and downstream demand can substantially limit these benefits. Coordinated deployment of high-value utilization, expanded treatment capacity, and interprovincial transport could reduce these constraints, with cumulative cost savings of nearly CNY 4000 billion and GHG mitigation of 6 billion tons of CO₂eq by 2050. Spatial planning should therefore be considered explicitly in the design of battery circular-economy systems.

CRediT authorship contribution statement

Hongyi Xie: Conceptualization, Data curation, Formal analysis, Investigation, Methodology, Resources, Software, Validation, Visualization, Writing – original draft, Writing – review & editing. Huaqing Li: Data curation, Investigation, Resources, Writing – review & editing. Sijing Wang: Writing – review & editing. Zhixiu Han: Writing – review & editing. Bin Chen: Conceptualization, Writing – review & editing. Hanyun Lv: Data curation, Investigation, Resources. Yutao Wang: Conceptualization, Funding acquisition, Project administration, Resources, Supervision, Writing – review & editing.

Data and code availability statement

All data used for this study are provided or available from publicly accessible sources, as cited in this manuscript and Supplementary Information. The data and code used to produce the outputs presented in this paper can be accessed via GitHub (https://github.com/XieEco/EOL_EVBAT_CHINA).

Declaration of generative AI and AI-assisted technologies in the manuscript preparation process

During the preparation of this work, the authors used ChatGPT in order to improve language and readability. After using this tool/service, the authors reviewed and edited the content as needed and take full responsibility for the content of the published article.

Declaration of competing interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the Supplementary data to this article:

Additional results of the optimal solution (Supplementary Fig. S1), additional descriptions of literature review of retired battery recycling technologies (Supplementary Text S1) and power storage application policies (Supplementary Text S2), retired battery supply module (Supplementary Text S3), retired battery demand module (Supplementary Text S4), cost benefit analysis and Monte Carlo approach (Supplementary Text S5), existing metallurgical capacity (Supplementary Text S6), uncertainty analysis (Supplementary Text S7).