A crucial question in the era of climate change is whether artificial intelligence (AI) will tackle this existential danger already wreaking havoc on lives and livelihoods. The emerging reality is complicated. AI helps businesses run faster and smarter, but these efficiency gains come with the gigantic carbon footprint of heavy energy-consuming data centres. However, by learning from global experiences and enforcing meaningful regulations, Southeast Asia can become a leader in data centres while dealing with this climate conundrum.

On the benefit side of the AI equation, its systems can forecast floods, storms, heatwaves, and natural hazards more quickly and accurately. For example, in October 2024, meteorologists used AI weather models to accurately forecast that Hurricane Milton would land near Siesta Key, Florida. The capacity to assemble vast meteorological data at lightning speed is hugely valuable to Southeast Asia. The first AI-powered systems in the region are beginning to deliver weather predictions in the Philippines at 10 times higher resolution than before. 

Furthermore, AI-powered smart city technologies help make public transport systems work more smoothly. Waste management systems driven by AI help increase recycling rates. In the UK, for example, recycling company Recycleye uses AI to locate materials for sorting, lowering contamination rates and increasing recycling volumes. Singapore’s waste management integrates AI systems that use advanced sensors and machine-learning algorithms. For agriculture, AI-powered smart machines, robots, and sensors provide real-time monitoring of weather, soil conditions, and crop needs, leading to better water use.

But on the cost side, data centres massively use electricity, which remains predominantly generated by burning fossil fuels despite corporate strategies for increasing renewable energy for AI. Ever-expanding computer models handle voluminous data sets, with big heat-emitting data processors running day in and day out, often for several months. Maintaining optimal temperatures in data centres requires energy-intensive cooling systems (ChatGPT uses ten times more energy per search than a conventional Google search).

The International Energy Agency forecasts that electricity consumption from data centres for AI and cryptocurrency could increase by 117 per cent from 460 terawatt-hours in 2022 to more than 1,000 terawatt-hours in 2026. Put in perspective, these industries, which currently consume as much power as all of Germany, could, in four years, consume as much as all of present-day Russia. Southeast Asia, especially Singapore, Malaysia, and Indonesia, would lead the drive: one estimate is that energy demand by data centres will increase by 160 per cent by 2030 in the region.

Greener AI systems are urgently needed, with standard criteria to accurately measure their environmental impact. There must also be rigorous environmental regulation of AI and greater transparency from companies about their emissions. Tougher and targeted regulations will motivate the lowering of a model’s complexity without affecting performance. Diverse innovations offer learning experiences in energy efficiency, notably the latest models at DeepSeek, a Chinese AI start-up, which require 10 times less computing power than its competitors to generate comparable results.

…the region has a challenging but binding responsibility to pursue environmentally sustainable data centres.

One way to green the business is to choose locations near renewable energy sources or favourable climate conditions for extensive cooling systems. The design and operation of data centres can also incorporate natural ventilation, optimised thermal management, and improvements in server utilisation, cooling technologies, and energy-efficient hardware.

Advancements can build on progress in reducing energy overheads of AI computing, as evidenced by the halving of power usage effectiveness (PUE) in cutting-edge data centre facilities over the past 15 years in the US. Tech giants like Google and Microsoft are leading the way. The wide differences in levels suggest the scope for far-reaching change; for example, higher-PUE Singapore and Texas could look at the experiences of lower-PUE Netherlands and Sweden.

Southeast Asia has an opportunity to lead in the greening of data centres. In May 2024, Singapore launched its Green Data Centre Roadmap to add 300 megawatts of new data processing capacity by deploying energy-saving equipment. The following month, Thailand announced plans to attract data centre investments by expediting the development of the Direct Power Purchase Agreement for renewable energy. Malaysia has secured financing for two green data centres, further signalling Southeast Asia’s potential in this sector.

The region has unique advantages. In Southeast Asia, the growth of edge, cloud, hyperscale, or colocation — industry terms to indicate variations in physical location and data processing power of data centres — is partly due to the region’s rapidly expanding digital economy and Internet penetration, which raise the demand for data storage and processing capabilities. The rising adoption of cloud services also fuels the need for onshore data centres to support cloud infrastructure and services. Local data centres comply better with data sovereignty and data protection laws in Southeast Asia. Indonesia’s Personal Data Protection Law, for example, permits personal data to go outside the country only if security measures are met, which poses challenges for overseas data centres to have Indonesian clients.

Amid this potential and peril, Southeast Asia must persist with the daunting agenda of going green. With the high and rising share of fossil fuels in the energy mix and the already excessive carbon emissions, coupled with the high cost of its climate disasters, the region has a challenging but binding responsibility to pursue environmentally sustainable data centres.

At the end of the day, data centres must source energy from solar, wind, geothermal, and nuclear sources to drastically reduce their carbon footprint. All economic sectors face the cost of transition, but digitalisation initiatives that promise to make lives and livelihoods more sustainable are especially accountable to avoid adding to emissions and global warming. With its headway in data centres, Southeast Asia would do well to adopt global lessons and enforce economic regulations that promote the green economy.

2025/93