Southeast Asia’s technology economy is entering a phase where physical infrastructure matters almost as much as software. AI, cloud platforms and digital services all depend on computing capacity, driving rapid data centre growth. This shift is rapidly influencing investment, energy planning and the geography of technology talent.
Singapore remains a key data centre market, but Malaysia, Thailand and Indonesia are gaining importance as operators look for additional land, power and room to scale. As demand for AI applications accelerates, the region’s infrastructure decisions are increasingly shaping business strategies, investment flows and the future distribution of digital capabilities.
Kevin Choong from Panduit explores why Malaysia is becoming Southeast Asia’s new capital for AI data centres
Here are six ways the boom is reshaping Southeast Asia’s technology economy.
AI is turning compute capacity into a strategic resource
The biggest driver is AI. The training and running of AI models require far more computing power than conventional workloads. The global hyperscale demand is rising at a 14% compound annual rate. Additionally, technology companies plan on raising their AI capital expenditure by 61% in 2026.
Across Asia Pacific, data centre energy consumption has almost doubled between 2020 and 2024 and is likely to rise threefold within the next few years. On average, new data centres being built are exceeding 100 MW of power capacity, reflecting the scale of next-gen workload demands. This creates a different infrastructure equation for Southeast Asia, where reliable computing capacity is becoming increasingly important for AI startups.
Data centre growth is spreading beyond Singapore
Singapore remains a major Southeast Asian data centre market, but its land and power constraints are encouraging capacity to expand into neighbouring markets. Malaysia is a major beneficiary. In the first half of 2026, Southeast Asia accounted for around half of APAC’s under-construction data centre capacity, with Malaysia leading at 1,039MW and Thailand following at 859MW. Johor alone had 602MW under construction, while its total development pipeline reached 3,088MW.
Indonesia is also scaling rapidly. Jakarta’s development pipeline reached 1,699MW in H1 2026, a 56% increase, with 395MW under construction. Bangkok’s pipeline reached 2,084MW, almost doubling year on year. The result is a more distributed infrastructure map, with Singapore connected to Johor and Batam, while Bangkok and Jakarta develop regional roles.
Power is becoming one of the biggest technology constraints
The next data centre may be built where electricity can be secured, rather than where demand is strongest. Power availability has become a major constraint on data centre development across Asia Pacific, with growth shifting towards power-advantaged markets. The region recorded a record US$11.6 billion in data centre investment in 2025.
Data centres accounted for a record 9.3% of electricity consumption in Peninsular Malaysia during the second week of August 2026, compared with an average of 7% for the year. Separately, government projections indicate that data centres could account for as much as 31% of electricity consumption by 2035. Data centre development is increasingly part of the country’s energy planning. Access to grid capacity and long-term power arrangements is therefore becoming an important consideration in data centre development.
Cooling is becoming a core technology problem
Power is only part of the equation. AI chips generate significant heat, requiring more sophisticated cooling for high-density computing. AI-focused data centres can require more than twice the power density per server rack compared with traditional facilities. The resulting heat makes advanced cooling essential, while rising land, construction and equipment costs add to development complexity.
Southeast Asia’s tropical climate adds another consideration, as high temperatures can increase the energy required to cool data centre facilities. Reports indicate that in August 2026, high temperatures in Malaysia led to an increase in energy usage by data centres as they needed to cool down. Cooling is therefore becoming a performance issue as much as a sustainability issue. Liquid cooling and efficient facility design can determine how much computing capacity a site can support.
AI-focused cloud providers are creating a new layer of infrastructure
The growth of AI infrastructure has also accelerated interest in neocloud providers, which specialise in delivering GPU-focused computing resources. Unlike general-purpose cloud providers, neoclouds typically focus on specialised computing infrastructure, including GPU resources designed for AI and other high-performance workloads. According to CBRE, the category emerged in 2023-2024 and made a breakthrough in 2025 amid rising AI computing demand.
The rise of neoclouds such as CoreWeave and Nscale is described as part of the wider AI infrastructure boom. These companies provide specialised GPU cloud infrastructure to organisations that need large amounts of computing power without building their own facilities. For Southeast Asian startups, this could make advanced compute more accessible, allowing smaller companies to rent specialised GPU infrastructure rather than secure large blocks of capacity themselves.
The boom could reshape where startups and talent cluster
Data centres can influence the wider technology ecosystem by making more computing capacity available closer to customers. For Southeast Asian AI startups, this could support the development and deployment of applications that require lower latency or need to address data residency requirements. Infrastructure considerations are becoming increasingly important as startups move from experimentation towards production workloads.
The expansion of data centres could also influence technology talent. Operators require engineers, cybersecurity specialists and cloud architects, potentially creating demand for specialised skills in emerging data centre markets. However, the extent to which these locations develop broader technology ecosystems will depend on factors such as connectivity, access to talent and the presence of supporting businesses.
A more distributed tech economy
Southeast Asia’s data centre boom is about more than servers. AI is increasing demand for compute, while power and cooling are becoming critical engineering challenges. Malaysia, Thailand and Indonesia are adding capacity at a scale that is changing the infrastructure map. Neocloud providers are creating another layer between data centres and AI companies, while local compute could influence where startups and talent cluster. The next phase of Southeast Asia’s digital economy may therefore be shaped as much by power, cooling and computing capacity as by capital. As AI becomes embedded in more businesses, infrastructure-rich locations will become increasingly important.