Declining costs make solar plus storage economical for industrial captive power

Summary

Significant new coal-fired captive power generation is planned globally to support industrial applications. This is particularly true in Indonesia, where tens of GW of captive coal are in development to support metal processing at least in part for the manufacture of clean technologies such as batteries. However, recent cost declines in renewables and battery storage, along with performance improvements in battery cycle life, energy density, and critical-mineral intensity, have opened up new opportunities for economical electricity generation across the country. Here we assess the rapidly growing region of Kalimantan and show how, with targeted support policies, the co-location of solar PV generation with battery storage can achieve a Levelized Cost of Energy (LCOE) of 5.6–7.9 cents/kWh at present, varying with capacity factor and financing conditions, competitive with conventional captive coal generation for industrial applications. With a supportive policy and regulatory framework, the reduction of existing coal subsidies, and preferential financing terms, solar+storage can offer GW-scale potential as an economical alternative to new captive power plants, with a minimal land footprint, in Indonesia and other emerging economies.

Graphical abstract

Highlights

• •Solar+storage is cost-competitive with new captive coal power in Kalimantan, Indonesia
• •Co-location with solar cuts storage costs by up to 20% versus standalone deployment
• •Policy reforms could lower solar+storage LCOE to 5–7 ¢/kWh by 2030
• •Minimal land use: 0.1% of Kalimantan land could replace 4–6 GW captive fossil plants

Applied sciences; Energy systems; Engineering

Introduction

Whereas the decarbonization of electricity generation has traditionally focused on the bulk power system, the decarbonization of off-grid fossil fuel-fired power plants - otherwise known as “captive” plants - remains a key challenge. Globally, there are approximately 204 GW of captive coal and 353 GW of captive oil and gas plants in operation today (Figure 1A); these plants typically exclusively serve energy-intensive industrial users, with the metal processing and chemical industries constituting the largest shares (Figure 1B).^1,2 Historically, captive plants offered two primary advantages: affordability, generating electricity at a lower cost than electricity from the grid by sourcing their own fuel and avoiding interconnection-related tariffs, losses and fees; and reliability, by providing a constant supply of electricity particularly in regions which may be susceptible to power outages and/or power quality issues such as voltage fluctuations.^3,4,5 In recent years, however, increasingly volatile and higher fossil fuel prices, rising penetrations of utility-scale renewable energy (RE), and the scaling back of financing for new coal projects from major Asian banks are acting to reverse the economics of captive power.⁶ Further, captive plants are often subject to less stringent regulation and may not meet the latest efficiency standards, resulting in higher emissions per unit of electricity generated compared to grid-connected plants. Yet an additional 40 GW of captive coal and over 180 GW of captive oil and gas remain in development - including plants that have been announced, pre-permitted, permitted, or in construction - warranting the assessment of alternative approaches which could supplant these captive fossil plants while providing affordable and reliable electricity.^1,2 Captive plants may provide either solely electricity or both electricity and heat, yet considering that the decarbonization of industrial heat remains more complex and often application-specific, the decarbonization of the hundreds of GW of captive fossil plants that solely provide electricity are immediate candidates for near-term replacement with cleaner alternatives.

Figure 1.

Location of captive fossil plants

(A) By fuel type. Color refers to primary fuel type; circles refer to operating plants, while triangles refer to plants in development; symbol size corresponds to the plant’s electrical capacity (MW_e). “In development” refers to plants that have been announced, are in pre-permitting, have been permitted, or are in construction.

(B) By industry. Color refers to the primary industry that the captive plant supports; circles refer to operating plants, while triangles refer to plants in development; symbol size corresponds to the plant’s electrical capacity (MW_e). The “Other” category includes plants whose industry is unknown.

The challenge is particularly acute and ironic in Indonesia. Over the past decade, captive coal capacity has grown 8-fold, reaching 11 GW in 2023, with another 14 GW remaining proposed or in development.⁷ The majority of this captive coal capacity serves the metal processing industry - including nickel, cobalt and aluminum smelters - which have been promoted by the government alongside bans of unrefined metal ores in an effort to turn Indonesia into a manufacturing hub for batteries and electric vehicles.⁷ Existing regulation within the country prohibits the construction of new coal-fired power plants, yet omits captive plants, leaving a major regulatory loophole for their continued development.⁸ Meanwhile, the Just Energy Transition Partnership (JETP) - which mobilizes US$20 billion of funding to support Indonesia’s energy transition - focuses on power sector emissions, deferring plans for the decarbonization of captive plants to an unspecified date.^9,10,11 This approach differs drastically from the world’s largest host of captive coal capacity, China; since 2018, the Chinese government has strengthened regulation on captive plants, halting the development of new projects, and encouraged their grid integration and market-based dispatch, citing the hindrance captive plants can pose to effective electricity market functioning.^12,13

Recent declines in the global capital costs of battery storage, coupled with performance improvements, have opened up new prospects for solar+storage to serve as a competitive and zero-carbon supply of firm electricity generation (see Figure S1). While pack prices have declined by 40% in the past five years, cycle life has concurrently improved from approximately 5,000 cycles to 10,000 cycles.¹⁴ Simultaneously, improvements in energy density and reductions in the intensity of critical minerals have further acted to enhance battery performance and lower costs. However, leveraging low-cost storage hinges upon the accessibility of low-cost solar as well. At 8 cents/kWh, Indonesia currently has some of the highest solar prices in Asia.^14,15 In contrast, the cost of solar in India ranges 3–4 cents/kWh despite only 20% higher-quality solar resources (Figure S2).⁵ India has achieved these solar prices - amongst the lowest in the world - by capitalizing on global solar capital cost declines and instituting a supportive policy and regulatory framework; this has included transparent large-scale reverse auctions, proactive land acquisition and transmission connections, payment guarantees, preferential financing, and an eliminated or reduced domestic content requirement.^15,16 Meanwhile, Indonesia’s solar costs have not kept pace with global trends, suggesting that policy and institutional barriers rather than technical challenges hinder the more widespread deployment of renewable+storage systems.

A number of studies have explored the use of solar+storage to replace fossil fuel generation across various global contexts, with some focused on decarbonization pathways tailored to Indonesia. Ziegler et al. estimate that storage costs would have to drop below $20/kWh for solar and wind to provide cost-competitive baseload power; despite offering comprehensive temporal coverage, however, their analysis is spatially limited, focusing on just four US states.¹⁷ Another study by Peng et al. studies the planning and operational considerations of replacing operational coal power with wind and battery storage in a small remote grid in Fairbanks, Alaska; despite valuable insights on investment cost and capacity trade-offs, the generalizability of the study’s findings is limited due to Alaska’s substantial seasonal variation in renewable potential.¹⁸ Studies in the context of the Global South have focused on the major coal consumer of India, with Gangaopadhyay et al. and Mohan et al. exploring the reliability implications and necessary cost declines to for renewable+storage plants to support a decarbonized energy system.^19,20 Both these studies in India, along with the US emphasize the need for sustained declines in renewable and storage costs to phase out existing fossil fuel generation and support the transition to a low-carbon energy system. However, due to depreciated capital expenditures, this requires larger cost reductions than those needed for renewables+storage to outcompete new fossil fuel plants. Further, studies that have assessed decarbonization pathways in Indonesia specifically have overwhelmingly focused on the power sector, omitting both existing and new captive coal plants.^{21,22,23,24,25,26} Globally, studies which aim to address emissions from captive power focus on retrofitting existing captive fossil power plants with carbon capture and sequestration (CCS) technology and fail to capture the rapidly falling costs of renewables and energy storage technologies which can serve needs for new-build captive power.^5,27,28 As such, this article aims to address three critical aspects that have remained underexplored in previous studies: the feasibility of solar+storage to serve as an economical alternative to new captive coal plants; the unique operating needs of off-grid captive power plants, rather than on-grid systems; and the impacts of targeted support policies specifically in an Indonesian context. This is especially pertinent to ensuring consistency between Indonesia’s climate commitments and green technology aspirations.

In this article, we evaluate the feasibility of solar+storage to supplant new captive coal plants through the lens of Indonesia, and specifically, Kalimantan. Kalimantan is a rapidly growing region that is set to host the new capital city, Nusantara, as well as 70% of Indonesia’s planned additions of metal smelter capacity, which is to be supported by over 5 GW of planned captive coal additions.^7,29 We first assess Indonesia’s solar resources and show that its solar generation profile is highly consistent across all months of the year, with the average capacity factor hovering between 16 and 20%. We then compare various configuration options for solar+storage, comparing costs with conventional captive coal-fired generation. When solar is co-located with battery storage, we find that the resulting hybrid system is cost-competitive with new, subsidized coal plants, resulting in a levelized cost of energy (LCOE) of 5.6–7.9 cents/kWh at present that varies with capacity factor and financing conditions. By reducing coal subsidies to reflect the international coal price and providing concessional financing to new projects, solar+storage can supply firm power at a steep discount to new captive coal plants as well as natural gas-fired plants. With a conducive policy framework that supports cost reductions of solar+storage systems, we estimate a large, GW-scale potential for the deployment of solar+storage in Kalimantan with a minimal land footprint, along with other equatorial regions faced with the challenge of decarbonizing captive fossil plants.

Results

Kalimantan hosts high-quality solar resources

Indonesia has significant solar energy potential. As shown in Figure 2A, the photovoltaic power potential measures above 1000 kWh/kWp across most of the country, with the exception of some lesser-populated eastern areas. This corresponds to an annual solar capacity factor between 14 and 18%. In Kalimantan’s coastal regions in particular, where most industrial facilities and thus captive coal plants are to be located, the annual solar capacity factor averages upwards of 15%. While this is slightly lower than some of the world’s top solar regions, it is still sufficient for large-scale photovoltaic (PV) deployment, considering Kalimantan’s relatively stable weather patterns and low seasonal variability.

Figure 2.

Assessed solar resource potential in Indonesia

(A) Photovoltaic power potential across Indonesia.³⁰

(B) The average hourly solar PV output in Kalimantan, shown as a percentage of installed solar PV capacity for key months, remains consistent throughout the year.

We next further analyze the quality of the solar resource potential in East Kalimantan, the province where the majority of existing and in development captive coal plants are located within Kalimantan. We find that the average hourly output from solar PV in East Kalimantan hovers between 45 and 60% throughout the year, as shown in Figure 2B. Unlike several other regions of the world, which experience significant seasonal variation in hourly solar PV output, Indonesia’s solar generation profile remains highly consistent throughout the year as a result of the country’s close proximity to the equator. The absence of significant seasonal variation removes the need for additional seasonal storage, thereby reducing the capacity and thus the cost of the storage system that would otherwise be required to ensure a consistent output from a solar PV generation system throughout the year. This acts to improve the competitiveness of solar+storage as a zero-carbon solution for captive power supply in Indonesia, where a steady power output is required for many industrial applications, including metal processing, which may be limited in their ramping flexibility.

Economics of solar+storage

In order to serve as a baseload and generate a consistent power output over the course of the day to support the relevant industrial application, the solar farm must be co-located with an adequately sized battery storage system. Such co-location also enables significant cost savings resulting from the sharing of balance of system (BOS) costs compared to if the solar farm and battery system were sited separately, as shown in Figure 3A. The costs of the BOS for a co-located solar+storage system can be reduced by approximately 30–50% while the Engineering, Procurement and Construction (EPC) costs can also be reduced by up to 50%, compared to a standalone utility-scale battery storage system; the costs of the battery pack and the structural Balance of System (BOS) remain the same.³¹ With the GW-scale deployment of co-located systems and low battery pack prices, this could ultimately translate into a 30% reduction in the total capital cost of the battery system to approximately $157/kWh today, with further reductions in pack prices and electrical BOS costs to approximately $114/kWh by 2030. These prices are increasingly being achieved in some areas of the world today: for example, the latest standalone storage auction by Gujarat Urja Vikas Nigam Limited (GUVNL) in India in April of 2025 saw a winning bid of 2.8 lacs/MW/month without any subsidy, which reflects a capex of ∼$130/kWh and Levelized Cost of Storage (LCOS) of ∼$34/MWh.

Figure 3.

Costs of battery storage systems in Indonesia

(A) Co-location and achievement of low pack prices through at-scale deployment, a conducive policy framework, and the removal of certain key barriers such as transmission, land availability, and so forth can reduce capital costs of battery storage systems by up to $100 US$/kWh. For full assumptions, see Table S1.

(B) LCOE of GW-scale solar+storage plants compared to new coal in 2023 real US$ for a 65% system capacity factor.

(C) LCOE of GW-scale solar+storage plants compared to new coal in 2023 real US$ for a 100% system capacity factor. The LCOE assessment assumes a gas price of $9.0/MMBtu, an Indonesian coal price of $70/ton, and a battery pack price of $139/kWh.

We next estimate the evolution of the levelized cost of electricity (LCOE) of a solar+storage plant against a conventional coal-fired power plant as well as alternatives with natural gas, for both a 65% and 100% capacity factor, as shown in Figures 3B and 3C. We find that today, large-scale solar+storage deployment can already be competitive with existing coal-fired power plants in Kalimantan: with normal financing options, the LCOE ranges between 7.9 and 8.3 cents/kWh for both a 65% and 100% capacity factor, respectively, while preferential financing options reduce the LCOE by approximately 25% to 6 and 6.4 cents/kWh for the respective capacity factors. These electricity prices are competitive with the existing commercial and industrial tariffs in Central Kalimantan, a respective 7.7 and 7.8 cents/kWh in 2022.³² Meanwhile, the costs of solar+gas, although competitive with captive coal today, are set to increase in the future with the rising price of natural gas fuel, which Kalimantan would have to import in the form of LNG from other Indonesian regions due to the lack of regional production. The solar+storage+gas alternative remains more expensive, especially as the competitiveness of solar+storage is set to improve with technological improvement and further commercialization, with the LCOE falling to 5–7 cents/kWh by 2030 - varying with capacity factor and financing type.

In comparing solar+storage to coal, we analyzed two scenarios utilizing coal-fired capacity. The first scenario uses a coal price of $70/ton, which is the current price of coal sold to the Indonesian power sector as a result of heavy subsidization by the Indonesian government. However, this is only a quarter of the international benchmark price, which has averaged approximately $200/ton in recent years. In Figures 3B and 3C, we show how incorporating this international coal price into the analysis significantly raises the cost of coal-fired electricity. In addition, the true cost of new coal power plants also must include the opportunity cost of not being able to sell the coal at prices significantly higher than the price cap in the coal sector, suggesting that the subsidy on coal should be eliminated or at least reduced. With the cost of coal also projected to increase in the coming years, solar+storage emerges as a competitive option to serve baseload at a deep discount to coal.

We next perform a bottom-up optimization to minimize the total levelized costs of a solar+storage system, incorporating an hourly balancing that accounts for solar generation, storage charging and discharging, and any curtailment or shortage for every hour of the year. The respective operating profiles of the different configurations of solar+storage, optionally with natural gas as backup, are shown as an hourly average in Figure 4. We assume a maximum load served of 100 MW, although calculations were performed on a per MWh basis and can thus be scaled as desired. On average, we find that approximately 40–50% of daily solar energy needs to be stored in batteries or other forms of energy storage. Of all evaluated configurations, solar+storage can achieve a capacity factor of 95% at a cost of $92/MWh with 0 kg CO₂/kWh_el; configurations that include a natural gas turbine and achieve a 100% capacity factor are similar if not more expensive, with higher emissions per unit of generated electricity as well.

Figure 4.

Daily profiles of captive coal alternatives

Assumes a 100 MW maximum load served, based on 2025 costs in 2023 US$ with no preferential financing.

(A) Annual average of solar+storage at a 65% capacity factor.

(B) Annual average of solar+storage at a 100% capacity factor.

(C) Lowest RE day for solar+storage at a 100% capacity factor.

(D) Highest RE day for solar+storage at a 100% capacity factor.

(E) Annual average of solar+gas at a 100% capacity factor.

(F) Annual average of solar+storage+gas at a 100% capacity factor, with maximum solar curtailment set at 10%. Notes: For options with gas, the gas price is assumed to be $11/MMBtu. Only direct emissions at the generation site are included within the emissions intensity. Life cycle emissions, such as those associated with raw material extraction, manufacturing, or decommissioning, are beyond the defined system boundary.

While our primary emissions estimates reflect direct operational emissions, we note that life cycle considerations further reinforce the contrast between these different configurations. Solar+storage systems incur most of their associated emissions upfront during manufacturing and construction, whereas systems relying on natural gas generation emit continually throughout their operational lifetimes. Numerous high-profile studies have stressed that for power-sector planning and decarbonization timelines aligned with climate targets, operational emissions dominate cumulative impacts and are the most policy-relevant.^33,34 Even when life cycle emissions (including pre-, during, and post-operation) are considered, solar PV with battery storage typically emits 20–80 g CO₂e/kWh, representing significantly lower total emissions per unit of electricity generated than gas-based systems which typically range 400–1000 g CO₂e/kWh.³⁵ Thus, our focus on operational emissions is consistent with both the structure of most planning frameworks and the long-term emissions performance of the technologies studied.

We additionally run a low clean costs sensitivity to compare with the business-as-usual (BAU) cases described above, exploring different configurations of solar, battery storage, and natural gas along with normal financing terms. The low clean costs case reflects a solar capex of $400/kW and total battery capex (including pack, BOS and EPC) of $100/kWh, reflecting costs that are increasingly being achieved in some parts of the world. In other words, the low clean costs case reflects realistically achievable costs Indonesia were to implement a supportive regulatory framework to aid in low-cost, large-scale clean tech deployment. As seen in Table 1, the total LCOE of the low clean costs case is nearly 30% lower than in the BAU case, demonstrating the increasing competitiveness obtainable through conducive policies.

Table 1.

Results for the low clean costs case, versus the business-as-usual case

Scenario	Business-as-usual case	Low clean costs case
Solar+storage, CF = 95%	Solar LCOE: $60/MWh + Storage adder: $32/MWh = Total LCOE: $92/MWh	Solar LCOE: $45/MWh + Storage adder: $21/MWh = Total LCOE: $66/MWh (−28%)
Solar+storage+gas, CF = 100%	Solar LCOE: $33/MWh + Storage adder: $14/MWh + Gas adder: $47/MWh = Total LCOE: $94/MWh	Solar LCOE: $25/MWh + Storage adder: $9/MWh + Gas adder: $47/MWh = Total LCOE: $81/MWh (−14%)

Discussion

Recent declines in the costs of clean energy technologies, and in particular battery storage, along with performance improvements, have opened up new prospects for solar+storage to serve as a low-cost and zero-carbon supply of baseload electricity generation. This is a key opportunity for the hundreds of GW of operating and planned captive fossil plants globally today, and especially for Indonesia, where tens of GW of captive coal are in development to support metal processing for the manufacture of clean technologies. Here, we assess the potential for solar to be deployed competitively at scale and co-located with battery storage in the rapidly growing region of Kalimantan. We find that the region has high-quality solar resources, which, with targeted policy support, can be leveraged through the co-location of solar PV generation with battery storage. The resulting Levelized Cost of Energy (LCOE) is estimated to vary between 5.6 and 7.9 cents/kWh depending on the capacity factor and financing conditions, which is competitive with conventional captive coal generation. With the reduction of existing coal subsidies and concessional financing, solar+storage can serve as baseload at a steep discount to new captive coal-fired plants as well as captive natural gas-fired plants.

With ambitious climate targets and booming metal smelter growth, Kalimantan stands at a pivotal juncture in shaping the sustainability of its development trajectory. Solar PV and battery storage technologies now offer Kalimantan a competitive and clean alternative to the proposed captive coal power expansion. And although the economics of captive plants’ grid integration may vary widely by region, we show here how solar+storage systems could be deployed at scale to serve as zero-carbon, off-grid baseload. While we assume conservative battery pack prices of $100/kWh in 2025, pack prices are already available at $80/kWh in some regions of the world and are projected to decline further by 2030, improving project economics. In addition, solar plus storage-based energy systems, once deployed, will have nominally constant energy costs over the next 25–30 years, suggesting a decline in real terms. This stability offers substantial economic advantages, shielding the industry from the impacts of fuel price volatility and supply disruptions. The pursuit of this development also circumvents difficulty and uncertainty related to securing capital needed for new conventional captive projects, as many Asian banks pull financing for new coal plants. Further, the large-scale deployment of solar+storage systems, in lieu of conventional captive coal, facilitates consistency between Indonesia’s climate commitments and aspirations of developing integrated value chains of manufacturing green technologies.

If 20 GW of solar PV capacity is deployed in Kalimantan, it could supplant over 4–6 GW of captive coal or gas capacity operating at 65%–100% capacity factor. This would require about 0.4 million hectares of land, representing 0.1% of the land area of Kalimantan. Put differently, it is equivalent to a tiny fraction - about 0.3% - of the 12 million hectares dedicated to palm oil production in Indonesia, 90% of which appears to be in Kalimantan and Sumatra.^36,37

However, the actual scale and speed of deployment in the real world will hinge on the establishment of an enabling policy and regulatory framework. This involves the gradual elimination of current subsidies on coal for electric power generation, expediting the cost reduction of clean energy technologies—especially solar and battery storage—and eliminating obstacles to rapid deployment. Several relevant lessons could be adopted from other countries, particularly India, which has successfully implemented large-scale reverse auctions, volume aggregation, proactive transmission connections, preferential financing, payment guarantees, and incentivizing domestic manufacturing to accelerate the deployment of wind and solar power, while achieving dramatic reductions in costs. For example, between 2010 and 2022, solar PPA prices in India dropped by 85% from 15 cents/kWh to under 3 cents/kWh - the largest reduction for any country - propelled by the innovative policy and regulatory framework as well as the global trends. If similar policies are implemented in Indonesia, the cost of solar generation can be further reduced up to approximately 4 cents/kWh, critical for providing a clean and cost-effective alternative for the industry and placing Kalimantan at the forefront of Indonesia’s energy transition. The sharing of best practices beyond India and Indonesia across other emerging economies - and particularly those located in equatorial regions - can help deploy solar+storage as a cost-effective alternative to both existing and planned captive fossil plants.

Limitations of the study

This study is subject to several methodological and data-related limitations. First, the solar resource profile used in the optimization is derived from MERRA and MERRA-2 data,³⁸ which, despite being widely used, has been noted to exhibit biases in renewable resource estimation.²⁰ However, these issues tend to be more prolific with wind rather than solar, where the 50 km spatial resolution of MERRA-2 data may be too coarse to accurately represent localized wind resources and thus contribute to overestimating wind speeds. Second, the nascent state of large-scale solar PV deployment in Kalimantan and in Indonesia more broadly constrains our ability to verify the accuracy of the solar profile used in this analysis against ground measurements. In the case that the actual resource quality proves lower than modeled, additional solar PV capacity would be required, likely increasing the levelized cost only marginally. Third, the techno-economic results rely on current and projected cost benchmarks that may evolve differently under future market and policy conditions. Finally, while the study focuses on operational emissions, life cycle impacts—though significantly lower for solar+storage than for fossil alternatives—were not modeled in full detail, and future work could incorporate a broader set of environmental and socio-economic factors.

Resource availability

Lead contact

Further information and requests should be directed to the lead contact and corresponding author, Nikit Abhyankar.

Materials availability

The study did not generate new materials.

Data and code availability

• •All data reported in this article will be shared by the lead contact upon request.
• •This article does not report original code.
• •Any additional information required to reanalyze the data reported in this article is available from the lead contact upon request.

Acknowledgments

The authors would like to thank local stakeholders and industry partners for valuable discussions and insights that contributed to the development and refinement of this work.

Author contributions

E.C., N.A., U.P., and A.P. designed research; E.C. and N.A. performed research, analyzed data, and wrote the article.

Declaration of interests

The authors have no interests to declare.

STAR★Methods

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Other
Solar, coal, gas, & battery pack capital costs	Data Sets14	–
O&M, EPC, BOS costs	Annual Technology Baseline (ATB)39	https://atb.nrel.gov/electricity/2022/data
Solar resource profile	Pfenninger and Staffell38	https://doi.org/10.1016/j.energy.2016.08.060
Cost of equity & debt	Data Sets14, Lee et al.40	https://www.nrel.gov/docs/fy19osti/71814.pdf

Method details

Resource assessment

We first assess the solar resource potential in Kalimantan and specifically at the Kalimantan Industrial Park (KIP) which is located at approximately (2.5, 117.9). We use Renewables.ninja to obtain the 8760 solar profiles, which utilizes a dataset based on MERRA and MERRA-2 with a spatial resolution of 1/2° latitude and 2/3° longitude.³⁸ We construct a grid of distinct sites around the KIP, ranging from 1.5 to 3.5 degrees in latitude (in 0.5 degree increments) and 115 to 117.9 degrees of longitude (in 0.66 degree increments). This translates to 15 distinct onshore sites. The solar profiles from these sites are averaged on an hourly basis to obtain the ultimate solar profile for the bottom-up system optimization.

Technology costs

The capex of coal and natural gas as well as fuel costs are based on Indonesia-specific values.¹⁴ The price of Indonesian coal is set to increase by 2% per year.

Solar energy prices in Indonesia are significantly higher than other emerging economies like India, despite both countries having similar solar resource potential (see the supplemental information for details). According to BNEF,¹⁴ Indonesia had an average utility-scale solar photovoltaic price of $80-100/MWh in 2023, while India had an average price of $30/MWh. However, with large-scale deployment, supportive policies and regulations, well-designed auctions, and removal of key barriers, Indonesia could achieve solar energy prices on par with India’s. Barriers holding back cost reductions include lack of transmission infrastructure, land availability issues, and high domestic content requirements.⁴¹ For example, India has been able to lower the solar costs from $80/MWh in 2015 to $30/MWh by 2019 through transparent auctions, guaranteed power purchase agreements, facilitative policies, and rapid capacity growth (Abhyankar et al., 2021). We believe that adopting similar best practices could allow Indonesia to achieve globally competitive PPA prices as well. Therefore, for this analysis, we set Indonesia’s solar PV capital cost to be similar to India’s, and decrease by 2% per year, reflecting forecasted trends.⁴²

Similarly, for estimating battery storage costs we use global benchmarks on pack prices and National Renewable Energy Laboratory’s (NREL) Annual Technology Baseline (ATB) for 2022 for the balance of system (BOS) and engineering, procurement and construction (EPC) costs.^14,39 Although battery pack prices have fallen to around $80/kWh in some areas, we conservatively assume a higher pack price at present in 2025 and match 2030 cost projections to the auction prices achieved in India.

The final Levelized Cost of Electricity (LCOE) of the respective hybrid system (solar+storage, with gas, or coal only) is calculated using an economic lifetime of 30 years and is taken to be the sum of the annualized capital expenses, operating expenses (non-inclusive of battery pack replacement given the high cycling life) and the marginal cost of fuel (for coal and gas-fired plants):

$$LCOE=\frac{{capex}_{solar,annualized}+{opex}_{solar}}{{CF}_{solar}·8760}+\frac{{capex}_{storage,annualized}+{opex}_{storage}}{{utilization}_{storage}·8760}+\frac{{capex}_{fossil,annualized}+{opex}_{fossil}}{{CF}_{fossil}·8760}$$

Where utilization_storage refers to the annual average throughput of the storage system divided by rated power. We assume a WACC of 10% in the normal financing case and a WACC of 7% in the preferential financing case. See Table S1 for full cost assumptions.

Bottom-up system optimization

We conduct a bottom-up system optimization to determine the cost-effective configuration of different system combinations: solar+storage at a 65% capacity factor (CF), solar+storage at a 95% CF, solar+gas at a 100% CF, and solar+storage+gas at a 100% CF. Using the averaged 8760 solar profile for the KIP, the optimization minimizes the project LCOE while incorporating an hourly balancing that accounts, on an hourly basis, for solar generation, storage charging/discharging, possible solar curtailment, and possible shortfall:

$$\mathit{min}\sum _{t=1}^{8760}{LCOE}_{solar}+{LCOE}_{storage}+{LCOE}_{fossil}$$

Subject to the following constraints…

Energy balance: $P_t^{solar}+P_t^{storage,discharge}+P_t^{fossil}=E_t^{served}+P_t^{storage,charge}\forall t$

Storage dynamics: $E_t^{storage}=E_{t-1}^{storage}+P_t^{storage,charge}-P_t^{storage,discharge}\forall t$

Where P variables refer to the power input/output in a given hour, $E_t^{served}$ refers to the total load served in each hour t, and $E_t^{storage}$ refers to the state of charge. In the solar+storage cases, to ensure sufficiently high renewable energy penetration, we impose a constraint that annual solar and storage generation must achieve the desired minimum capacity factor of 65% or 95%.

Published: October 15, 2025