Coal flexibility risks locking in gas in Peninsular Malaysia

Summary

The potential of coal flexibility to support a clean transition is highly system- and context-dependent. TransitionZero’s simulation of Peninsular Malaysia’s 2030 grid shows that between 69% to 90% of reduced coal output risks being replaced by gas, as solar capacity is already fully utilised.

In Sarawak, however, coal flexibility could enable low-cost coal-to-clean switching. Hydropower fully replaces reduced coal generation, with system costs increasing by less than 1%.

To implement coal flexibility effectively, countries need strong policy guardrails. Clear eligibility criteria, fixed sunset dates, emissions benchmarks, transparency requirements, and declining financial incentives are needed to ensure that coal flexibility supports — rather than delays — the coal-to-clean transition.

Coal flexibility: a system-level consideration

For coal flexibility to play a constructive role in the clean energy transition, its impacts must first be assessed at the system level.

As a case study, TransitionZero developed a model of Malaysia’s 2030 power system to evaluate whether and how flexible operation of existing coal plants could enable higher levels of solar integration, and at what cost.

Malaysia was selected because coal is expected to remain a significant part of the power mix for the foreseeable future, despite the country’s substantial solar potential and opportunities for renewables to meet demand growth and replace ageing fossil generators.

Malaysia hosts the third-largest coal power fleet in Southeast Asia, with 13.3 GW of installed capacity. Although the government has committed to halting new coal development and retiring existing assets as their offtake contracts expire, the final coal units — with a combined capacity of 2.6 GW — are still scheduled to operate until 2044 under the current phase-out roadmap.

Coal utilisation rates are also high. Between 2020 and 2023, average annual capacity factors ranged from 70% to 79%, with several large units reaching 85-86% in some years, levels notably higher than those observed in regional systems. This means that coal’s current and future emissions footprint remains a key challenge for Malaysia’s power sector decarbonisation.

Malaysia’s coal assets are distributed across two independent grid systems that differ considerably in structure and resource endowment. Peninsular Malaysia relies heavily on coal and gas, which together supply around 94% of total generation. Sarawak, by contrast, is endowed with extensive renewable resources and currently sources 74% of its electricity from hydropower, making it arguably the cleanest grid in Southeast Asia.

These contrasts mean that the system-level impact of coal flexibility may differ markedly between the two grids, offering valuable insights into the conditions under which this solution may — or may not — deliver meaningful climate and cost benefits.

Model set-up

We developed a representative model of Malaysia’s 2030 power system using the open source Python for Power System Analysis (PyPSA) framework. Malaysia is represented as three nodes corresponding to its grid regions: Peninsular Malaysia, Sarawak, and Sabah. Peninsular Malaysia is interconnected with Singapore and southern Thailand, while Sarawak can trade power with Sabah and Indonesia’s Kalimantan.

The model operates at hourly resolution, representing each of the 8,760 hours in a year. The power system is represented across multiple generation technologies, including both thermal and renewable sources. Coal plants are modelled using a unit commitment framework, allowing the model to capture key operational and financial characteristics of thermal generation, such as start-up cost or minimum uptime/downtime.

While hourly resolution does not capture minute-level ramping behaviour typically required for VRE-compatible coal flexibility, it is sufficient to address the central research question: whether flexible coal operations can enable higher VRE penetration, and at what cost.

We modelled one reference scenario and three coal flexibility scenarios. The model year 2030 was selected to capture the near-term transition potential of coal flexibility rather than outcomes in a more distant timeframe.

• Reference Scenario: this baseline scenario assumes no constraint on coal dispatch.
• Flex10, Flex20, and Flex30: in these scenarios, annual coal utilisation is reduced by 10%, 20%, and 30%, respectively, relative to the Reference Scenario baseline (87% in Peninsular Malaysia, and 62% in Sarawak).

Across all scenarios, a no-new-coal constraint and a policy solar target of 7 GW nationwide were implemented, in line with the National Energy Transition Roadmap. We also assumed the scheduled retirements by 2030 of three power stations, namely Sultan Aziz (1,600 MW), TNB Janamanjung GF1 (2,070 MW), and Sekingkat (210 MW). As a result, installed coal capacity for the full year 2030 is capped at 8,510 MW in Peninsular Malaysia, and 894 MW in Sarawak.

For details on the input data and model results, please visit our documentation here.

Coal flexibility risks up to 90% gas switching, if not supported by sizeable clean energy

The outcomes under these assumptions yield the following insights.

In Peninsular Malaysia, reduced coal output are backfilled by gas by up to 90%, as solar is fully utilised

In the main grid, reductions in coal dispatch primarily result in coal-to-gas substitution rather than increased renewable generation. Across the Flex10, Flex20, and Flex30 scenarios, between 69% and 90% of reduced coal output is replaced by additional gas generation at the annual level. Solar generation remains unchanged because the system builds and utilises the maximum modelled solar capacity of 26 GW. With solar already fully deployed and dispatched, no additional renewable generation is available to cover the curtailed coal.

An illustrative intra-day profile on 1 June 2030 highlights these dynamics.

The simulation indicates that reducing coal dispatch alone has limited impact in driving a coal-to-clean transition in Peninsular Malaysia. Rapid electricity demand growth and the fossil fuel–heavy supply mix mean that lower coal output is largely offset by higher gas generation. This underscores the need for more ambitious solar targets and faster deployment of clean sources of supply in the coming years.

In terms of system costs, the Flex10 scenario yields modest but positive savings of US$16.7 million relative to the Reference Scenario. However, deeper coal curtailment increases system costs as gas generation rises, with total cost increasing by US$81.7 - 192.1 million in the Flex20 and Flex30 scenarios. This corresponds to an abatement cost range of US$9.5 - 15.4 per tonne of CO2.

In Sarawak, renewables displace coal at less than 1% system cost increase

Sarawak presents a markedly different outcome to the main grid. Because hydropower already supplies the majority of electricity and coal accounts for only 10% of total system capacity in 2030, reductions in coal output are almost fully replaced by additional renewable generation.

Across all flexibility scenarios, displaced coal generation is met by increased hydropower output, demonstrating the region’s strong potential for coal-to-clean switching.

The intra-day dispatch pattern on 9 September 2030 illustrates this clearly.

The system cost impact of coal flexibility in Sarawak is also limited. Total system cost increases by just 0.5% in Flex10, 0.7% in Flex20, and 1.0% in Flex30 relative to the Reference Scenario. The resulting abatement costs range from US$10.9 to US$13.2 per tonne of avoided CO2, with the lowest observed in the Flex30 scenario.

Policy guardrails for effective implementation

The Malaysian simulation underscores that the effectiveness of coal flexibility in supporting a clean transition is highly system- and context-dependent.

Regions with abundant clean energy resources and limited reliance on coal, such as Sarawak, can reduce emissions by operating coal units more flexibly. In contrast, regions where clean alternatives remain constrained relative to demand growth, such as Peninsular Malaysia, may see little coal-to-clean substitution unless coal flexibility measures are paired with clear and binding renewables mandates.

To ensure that coal flexibility supports — rather the delays — the coal-to-clean transition in Southeast Asia, national programmes and pilots should incorporate clearly defined policy guardrails. These safeguards help align coal flexibility initiatives with national decarbonisation and no-new-coal commitments while providing clarity for international financiers evaluating transitional support. Without these, coal flexibility risks extending asset lifetimes, diverting scarce capital away from clean energy, and delaying renewable deployment.

The following guardrails should be considered from the outset in policy, market design, and contractual frameworks.

Selected participation with strict eligibility criteria

Coal flexibility should be pursued only where it delivers clear system value. Screening criteria may include technical suitability (e.g. plant age, boiler technology), efficiency at part-load operation, proximity to renewable sources and grid balancing needs, remaining economic life, projected emissions, and alignment with national coal phase-out and just transition plans.

These criteria should be transparent and periodically reviewed to reflect evolving system conditions, such as new renewable capacity or grid reinforcements.

Sunset provisions

All flexibility arrangements must include fixed, legally enforceable expiry dates that take into consideration timelines for coal-to-clean transition as backed by climate science. These sunset dates should be embedded in power purchase agreements, market participation agreements or regulatory mandates, and linked directly to a plant’s decommissioning schedule.

Once agreed, the retirement dates should not be renegotiable except to bring them forward.

Emissions measures and benchmarks

To ensure flexible operation remains consistent with climate objectives, programmes should impose a declining cap on operational emissions intensity (measured in gCO2 per kWh generated) and implement other disincentives such as carbon pricing.

These benchmarks should tighten over time to reflect the increasing availability of zero-carbon flexibility options.

A phased tightening of benchmarks could also guide a payment schedule that is not front-loaded, thereby rewarding continued emissions reduction and encouraging a gradual shift towards clean flexibility sources as the system evolves.

Transparency and reporting from a verified baseline

Robust monitoring, reporting, and verification frameworks must be in place to track operational performance and emissions outcomes.

Participating plants should publish annual reports detailing usage rates, ramp rates achieved, minimum stable generation levels reached and emissions intensity. These reports should be independently verified and made publicly available to ensure accountability.

Phase-down of financial incentives

Where flexibility-related payments or subsidies are provided, they should be explicitly time-limited and designed to decline over the course of the programme. This could involve a performance-based payment structure where disbursements are linked to the achievement of pre-agreed milestones.

Aligning payments with the build-out of renewables, storage and transmission capacity helps avoid long-term dependencies and ensures that scarce public finance is progressively redirected towards permanent clean energy solutions.

Where plants fail to meet agreed flexibility or emissions targets, revenue payments should be withheld or clawed back to preserve accountability and credibility.

Enhanced renewable energy targets

To reduce the risk that coal flexibility displaces clean energy, national energy plans should explicitly link coal flexibility programmes to near-term renewable deployment targets.

As renewables expand, the role of coal in providing flexibility should diminish accordingly. Policymakers can signal this by publishing guidance that flexibility payments are available only for specific services such as ramping or reserves, and that the scale of support will decline as renewable milestones are met.

This approach maintains incentives for operational flexibility while reinforcing renewables as the primary source of the future system. It also strengthens investor confidence that coal flexibility is a transitional measure directly tied to the pace of clean energy deployment.

Note: The model was developed in June 2025 and thus did not incorporate policy and market developments after that time, including but not limited to the Sarawak government’s solar deployment target of 1,500 MW by 2030.

This is the final in a three-part series exploring the potential of coal flexibility in Southeast Asia. Part I describes the ins and outs of coal flex, with an eye to Southeast Asia, and Part II analyses the system, operational, and contractual challenges to deploying coal flexibility in the region.

This blog was co-authored by Thu Vu and Isabella Suarez. Technical modelling was conducted by Handriyanti Diah Puspitarini, Dan Welsby, and Abhishek Shivakumar.