The recently-announced reform of Peninsular Malaysia’s electricity tariff mechanism is critical in reflecting the true costs of fossil fuels paid by heavy energy users such as data centres, according to the country’s energy regulator.
Siti Safinah binti Salleh, chief executive officer of the Energy Commission, said that the new mechanism, which separates energy, capacity, network and retail charges, will make it easier for companies to decide on how they can meet their clean energy commitments.
“If all the costs are bundled, no one knows how much you are spending on generation costs, of which 70 per cent is actually fossil fuel costs,” she said last week at the Malaysia edition of the Unlocking capital for sustainability forum.
“It was important to get companies to understand that if they invest in clean energy solutions – solar, batteries and what not – and build their own energy portfolio, then they know how much they are paying in comparison to fossil fuels,” Safinah added.
Announced in late June and effective from July 2025 until December 2027, the new tariff mechanism introduced the Automatic Fuel Adjustment (AFA) scheme, which will reflect changes in the market price of fuel and foreign currency exchange rates on a monthly basis.
Under the previous tariff structure, changes in fuel prices were only passed on to consumers every six months via the Imbalance Cost Pass-Through (ICPT) mechanism.
The advantage of the new structure is that companies will have to be sensitive to the price volatility of fossil fuels such as gas and coal, said Safinah.
Although investing in new solar power and battery storage systems might require high upfront investments, these are fixed costs that would reduce businesses’ exposure to volatile gas prices, she said.
“Restructuring the tariff was a very important step to ensure that the foundation for energy economics is right,” Safinah said.
Data centre challenges
The new tariff structure also categorises users into low, medium and high voltage, with heavier industrial users facing higher energy, capacity, network and retail charges.
Data centres, which are categorised as “ultra-high voltage”, face an expected 15 per cent increase in energy costs, analysts have said.
Reik Ong, founder and managing director of Saxon Renewables, said that policy changes have been one of the biggest external challenges for data centre players in the region. Saxon helps data centre developers source renewable energy across Southeast Asian markets.
Although news headlines have suggested that Malaysia’s data centre industry has been booming in recent years, Ong said that the reality on the ground has not been as positive. Changes in tariff rates mean that companies have had to reassess the financial viability of projects, especially as agreements to build and operate data centres often span 20 years or more.
“We have seen some companies here starting to sell their data centres because they were not able to source a buyer,” he said, speaking on the same panel as Safinah.
On top of that, the internal policies of technology giants on sourcing clean energy have evolved.
Reik Ong, founder and managing director of Saxon Renewables said that internal and external policy changes on renewable energy have been among the biggest challenges for the region’s data centre players. Image: Eco-Business
“For example, we recently spoke to some hyperscale data centre owners like Microsoft, who mentioned that they no longer accept unbundled renewable energy certificates (RECs),” said Ong. Unbundled RECs refer to certificates that are separate from their clean energy sources, which means that the companies purchasing these RECs may not be receiving and using the clean energy they have paid for.
But Ong welcomed the new AFA scheme as an improvement over the previous ICPT mechanism, and highlighted the raft of policies the Malaysian government has tabled in recent years to improve corporate access to renewable energy.
Malaysia’s policymakers have been liberalising the country’s power system over the past, enabling third-party access to the electricity grid through programmes such as the Corporate Renewable Energy Supply Scheme (CRESS). Launched in September 2024, it allows companies to buy power directly from renewable energy producers instead of relying on unbundled RECs.
Using AI for predictive analytics
Safinah acknowledged the complex needs of companies, many of which have different internal power procurement policies. To address this, she said that the Energy Commission has been engaging data centre companies to better understand their energy needs.
The commission has learnt, for instance, that data centres do not consume a fixed level of electricity throughout the day – consumption increases during lunch hours or at night, when individuals are off of work and more actively using applications such as TikTok or shopping online. The increase in the use of electric vehicles (EVs) is also reshaping the demand profile of Malaysia’s electricity grid, Safinah said.
“There is another category of data centres for artificial intelligence (AI) – these use 10 times more energy during their machine learning process. After that, it’s actually quite minimal when it comes to repeating the same answer,” she said.
But AI can also improve demand forecasting through enhanced data analytics, and even support predictive analysis on the supply side of renewable energy, said Safinah. For instance, the technology could be used to analyse factors such as cloud cover, which affect the solar power generation.
“When it comes to the energy transition, it is very important to (use AI to) reduce the uncertainty and enhance predictability so that we can ensure we are planning ahead,” she said. “We need to be able to forecast at what points of time there will be rain showers or storms and so forth, which will impact the system as a whole.”
“It’s not just about planning on a long term basis, which is what (is important for) investments, but planning operationally to ensure that we know what is going to happen throughout the day,” she added.
Calvin Woo, director at the Malaysia Centre for Fourth Industrial Revolution (MyDigital Corporation), said that AI is already being applied in the energy industry via companies like Malaysia’s national oil firm Petronas, which has used AI for predictive maintenance and energy management systems in its downstream-focused Pengerang Integrated Petroleum Complex.
Under its Transitioning Industrial Cluster initiative, MyDigital, a government agency established in collaboration with the World Economic Forum, has also worked with the Economic Planning Unit of Sarawak to get 18 organisations and institutions to collectively agree to reduce 21.35 billion metric tonnes of carbon emissions by 2040 in Bintulu, Sarawak. Signatories include Petronas, Sarawak state oil firm Petronas, Thai oil company PTTEP, steelmaker Press Metal, and petrochemicals giant Neste.
On the data centre front, Woo pointed out that Malaysia is in the process of crafting the Asean Sustainable Data Centre Guide to tackle emissions at an industrial level.
Last week, Malaysia’s Minister for International Trade and Investment (MITI) Tengku Zafrul Abdul Aziz said that Malaysia’s own framework for sustainable data centres will be introduced in October. While the framework will come under the purview of the Ministry of Digital, MITI’s Malaysian Investment Development Authority (MIDA) will serve as the focal point for all applications of new data centre projects and expansions.
