However, a recent report by global think tank the International Data Center Authority (IDCA) shines a light on some of the challenges developers face when looking to build new data centres in the region – challenges that must be navigated to realise the significant investment opportunities arising.
State of play in the market
There has been an explosion of AI products and services – tools that write, generate images, answer questions, and automate tasks that used to require human effort. Those tools run on computing power, lots of it. That computing power lives in data centres, often in large buildings packed with servers and networking equipment that run continuously, around the clock, consuming vast quantities of electricity in the process. As demand for AI has grown at pace, the race to build the infrastructure that supports it has intensified dramatically. Asia Pacific, and the Southeast Asian sub-region in particular, has found itself at the centre of that race.
The region has genuine appeal as a destination for this kind of investment. It has a large and growing digital economy, a population that is adopting technology at remarkable speed, and a geographic position that makes it a natural hub for international data connectivity. Governments across the region have, for the most part, been actively welcoming of data centre investment, offering incentives and streamlining approvals in ways that have made the region increasingly attractive to some of the world's largest technology companies.
One well-established city-state, Singapore, has served as one of the most prestigious and trusted data centre address in Southeast Asia, and it remains enormously important in that role. But it is also, essentially, full. Its government has taken a deliberate and disciplined approach to managing how much new data centre capacity it permits – any new facility must meet very high efficiency and sustainability standards, and approvals are carefully rationed. The result is a market with almost no spare space and a premium price tag on what little does exist. Singapore has, in a sense, chosen quality over quantity, positioning itself as the address of choice for the most demanding and compliance-sensitive workloads, rather than competing on sheer volume of capacity.
The natural beneficiary of these constraints has been the immediate geographical neighbours of this city-state. A few years ago, the areas just across Singapore’s borders barely registered as data centre markets. Today, Malaysia is among the most dynamic data centre markets in the entire Asia Pacific region. Land is available and affordable. Electricity is cheaper. Local governments have been enthusiastic and cooperative. Further, its proximity to the established regional hub of Singapore – close enough to benefit from its connectivity and its role as a business centre – makes these neighbouring locations an attractive alternative for development. This is reflected by the volume of projects either under construction or in active planning.
Other markets across Southeast Asia have also been emerging beyond merely peripheral destinations for data centre development. Thailand and Indonesia have moved with impressive speed, assembling attractive packages of government incentives and designating specific zones for technology infrastructure development that have drawn significant investment commitments from some of the world's largest technology companies. The investment opportunity in these markets across Southeast Asia is underpinned by strong domestic digital economies, with large internet-using populations and business communities adopting AI and cloud services at a rapid pace.
The challenges facing developers
Modern AI-grade data centres are extraordinarily power-hungry. They run continuously, at very high intensity, and the cumulative demand they place on national electricity grids is significant in a way that is genuinely testing the limits of infrastructure that was never designed with this kind of load in mind.
According to the IDCA’s report, the approximately 10,000 data centres in the world consume 2% of the world’s electricity. However, there are significant regional disparities, creating pressure points in specific national markets. Of the 67.7 GW of power estimated to be required to run the world’s data centres, 16.7 GW is needed in Asia Pacific. In Singapore specifically, 19.5% of the country’s total electricity consumption is linked to data centre operations – by far the highest proportion of any country in the world. The comparative figure in the US is 6%.
Governments across the region are responding – accelerating renewable energy programmes, building new generation capacity and, perhaps most strikingly, revisiting nuclear power as a serious option after years of limited interest. Whether these efforts can deliver sufficient capacity at the speed required to keep pace with the construction boom is one of the most consequential open questions hanging over the entire sector.
Closely related to the energy challenge, AI servers generate far more heat than conventional computing equipment, and they do so at a density and intensity that traditional air conditioning systems simply cannot manage effectively – particularly in a tropical climate where outdoor temperatures are very warm and humid year-round. The industry has converged on liquid cooling as the solution: systems that run chilled water or engineered coolant in close proximity to server components, drawing heat away far more efficiently than moving air can achieve. It works considerably better than air cooling for these workloads, but it costs meaningfully more to build, and it introduces engineering complexity that plays into project economics.
Water is another dimension of the sustainability story. Data centres use water – often in substantial quantities – primarily in their cooling operations, and in parts of the region where freshwater availability is already under pressure from urban growth and seasonal variability. This is becoming a genuine planning and regulatory consideration.
In at least some markets across Southeast Asia, data centre projects are already being scrutinised for their approach to water consumption, and there are signs that facilities which cannot demonstrate responsible usage are finding the path to approval more difficult than they might have anticipated. Singapore is one country the IDCA flagged as having put in place significant conditions on data centre development. Water stress is one of the factors for this: according to the IDCA’s report, Singapore scores 73 out of 100 in the IDCA’s water stress index, making it one of the top 10 countries in the world where hypothetical constraints on access to water could serve as a barrier to data centre development.
The regulatory picture across the region adds yet another layer of complexity. Southeast Asia is emphatically not a single, uniform operating environment. Different markets have taken very different approaches to governing data, technology and AI. Vietnam recently enacted standalone AI legislation, complete with a risk-based classification system for AI products and a requirement that certain categories of data be stored within national borders. Other markets, like Singapore, take a more flexible, principles-based approach and impose no such blanket obligations.
Geopolitics has also become a tangible concern in the data centre context. The world's major technology powers are competing for presence across the region's digital infrastructure. Technology enterprises with very different backgrounds and governance relationships are building facilities within the same markets. Questions about which technology is being used, and who has access to what data, are live considerations that are attracting attention at a governmental level across the region.
Taken together, what emerges is a picture of a sector in the middle of a boom, but which is also navigating a set of challenges around energy, water, regulation, and broader global dynamics.
Within the substantial complexity that must be navigated are significant opportunities for investors and developers.