Southeast Asia now hosts more than 2,000 data centres across Indonesia, Malaysia, Singapore, Thailand, Vietnam and the Philippines (Ember, 2026), with hundreds more under construction and over a thousand in planning. Data centre investment in the region could reach $30 billion by 2030 (Turner & Townsend, 2026), and demand is expected to grow at 20% annually through 2028, according to the U.S.-ASEAN Business Council. This isn’t a boom driven by cloud migration alone. It’s an infrastructure race shaped by AI workloads, sovereign data mandates, and a geopolitical tug-of-war between US and Chinese technology stacks.
What makes Southeast Asia’s data centre expansion different from what’s happening in the US or Europe is the speed and the constraints. The region is building AI-grade infrastructure in a tropical climate where ambient temperatures sit between 27°C and 35°C year-round, power grids weren’t designed for hyperscale loads, and multiple governments are simultaneously trying to attract investment while managing resources they don’t yet have in sufficient quantity.
Here’s how the buildout is unfolding, country by country.
How big is the data centre market in Singapore?
Singapore remains the region’s Tier 1 hub, with roughly 1GW of operational data centre capacity and a vacancy rate of just 1.4% (ARC Group, 2026). But the city-state’s approach is fundamentally different from its neighbours: constraint by design.
Singapore maintains strict controls on new data centre supply. Land is scarce, power is imported, and the government has imposed stringent green energy requirements on any new capacity. The result is a market where existing capacity commands premium pricing and new entrants face a genuinely difficult approval process.
Microsoft’s $5.5 billion commitment to Singapore — announced in early 2026 as part of a broader $6.5 billion Southeast Asian package — is the single largest tech infrastructure investment the country has ever received. The spend covers cloud and AI data centre infrastructure through 2028, alongside a programme providing free Copilot access to 200,000 students. Amazon Web Services committed $9 billion to expand its existing Singapore infrastructure in 2024, and Google Cloud operates an established cloud region in the country.
Singapore isn’t trying to be the region’s scale-out location. It’s positioning itself as the command node — the place where latency-sensitive financial and enterprise workloads run, where hyperscaler regional headquarters sit, and where regulatory frameworks are most mature. The heavy lifting, in terms of raw compute capacity, is increasingly happening across the Causeway.
Why is Malaysia becoming Southeast Asia’s AI compute hub?
Malaysia — and specifically the state of Johor, just across the border from Singapore — has emerged as the primary scale-out location for data centre expansion in the region. The combination of available land, relatively lower power costs, and proximity to Singapore’s connectivity infrastructure has made Johor a magnet for hyperscale investment.
The numbers are striking. Malaysia has more than 500 operational data centres. Roughly 300 are under construction. And around 1,140 are planned (Ember, 2026). The country’s Prime Minister confirmed in February 2026 that all new data centre applications unrelated to AI have been stopped — only projects demonstrating clear AI-related benefits receive approval.
YTL Power’s NVIDIA-backed AI campus, AirTrunk’s hyperscale facility, and Microsoft’s land acquisitions and planned cloud region in Johor all illustrate how concentrated the buildout has become. Google invested $2 billion in its first Malaysian data centre in 2024, and a sovereign cloud and AI services partnership with Dagang NeXchange Berhad followed.
Malaysia’s 2026 budget allocated approximately $490 million (MYR 2.1 billion) for a sovereign AI cloud, aiming to keep data and model training within national borders. The country plans to add up to eight gigawatts of gas-fired power by 2030 specifically to meet data centre demand. And Malaysia has revived its nuclear energy programme, setting a 2031 target for bringing atomic power online — driven in significant part by the electricity consumption of AI infrastructure.
The central challenge is energy. Fossil fuels generate 81% of Malaysia’s electricity, while solar and wind provide just 2% (Ember, 2026). A standard AI data centre consumes as much electricity as 100,000 households, according to the International Energy Agency. Malaysia’s data centre ambitions and its decarbonisation commitments are on a collision course that hasn’t been resolved yet.
What’s driving data centre investment in Indonesia?
Indonesia represents a different kind of opportunity. The country’s data centre growth is driven primarily by domestic demand — it has one of the largest internet populations in the region and a rapidly expanding digital economy. Investment has concentrated around Greater Jakarta, where connectivity, enterprise demand, and cloud adoption are strongest.
BDx Data Centers, headquartered in Singapore, operates six data centres in Indonesia, including a 500MW campus in West Java. Bridge Data Centres, backed by Bain Capital, has expanded across India, Malaysia, and Thailand from its Singapore base. The sheer scale of Indonesia’s population — over 280 million people — means that even modest per-capita cloud adoption generates massive aggregate demand.
Infrastructure constraints remain a real factor. Grid reliability, permitting timelines, and the country’s coal-heavy power mix have slowed some projects. But hyperscale providers continue to expand capacity in Jakarta and surrounding areas as demand for cloud and AI services grows.
How is Thailand positioning itself as a data centre destination?
Thailand’s data centre story has accelerated dramatically. The country’s total IT capacity grew by more than 2,000% between 2019 and 2024 (DC Byte), and capacity is expected to triple from 350MW in 2024 to roughly 1GW by 2027.
The headline investments tell the story. Microsoft committed over $1 billion for Thai AI data centre infrastructure through 2028. Amazon Web Services is investing $5 billion over 15 years. Google invested $1 billion, launching its Bangkok cloud region in January 2026. TikTok’s parent company ByteDance has pledged $8.8 billion for regional data centre development, with Thailand as a key focus. In the first half of 2025, Thailand’s Board of Investment approved data centre investments totalling THB 521.2 billion ($16.13 billion) from 28 projects.
The Eastern Economic Corridor — spanning three provinces east of Bangkok — has become the focal point for hyperscale construction. Chonburi province is hosting the largest upcoming projects, including DayOne’s 120MW Chonburi Tech Park and Bridge Data Centres’ planned 200MW campus. AI workloads already account for 28% of total Thai data centre capacity as of early 2025, up from 20% the previous year.
Chinese technology companies are matching Western investment levels in Thailand. Huawei, Alibaba, and Tencent have all established local operations, and Bangkok’s data centre market is becoming a genuine testbed for whether US and Chinese infrastructure stacks can coexist in the same country.
Thailand’s government is offering eight-year corporate income tax exemptions, import duty waivers, land ownership rights, and expedited work permits for foreign IT specialists through the Board of Investment. The country’s central location in Southeast Asia — enabling low-latency access to Vietnam and Malaysia — and multiple submarine cable landings make it increasingly competitive with Singapore on pure connectivity.
Where do Vietnam and the Philippines fit?
Vietnam and the Philippines represent the next tier of Southeast Asia’s data centre expansion. Both are attempting to attract hyperscale investment while addressing structural constraints around power, permitting, and regulatory readiness.
Vietnam’s position is distinctive because of its regulatory ambition. The country enacted its first standalone AI Law in December 2025, effective from March 2026 — making it one of the few nations anywhere with a comprehensive, dedicated AI governance framework. The law introduces risk-based AI system classification, a National AI Development Fund offering grants and financing to startups, and regulatory sandbox mechanisms for testing AI applications under relaxed compliance requirements.
Vietnam’s IT market is one of Asia’s most dynamic. The country’s national strategy targets 150,000 digital technology firms by 2035 and aims for the digital economy to account for at least 30% of GDP by 2030. Data centre investment is growing, though it remains smaller in absolute terms than Malaysia, Indonesia, or Thailand. Google, Microsoft, and Amazon have all signalled interest in Vietnamese cloud infrastructure, but committed spend has been more cautious than in neighbouring markets.
The Philippines faces similar dynamics — growing domestic demand, an increasingly tech-literate population, and government interest in attracting investment — but remains earlier in its data centre development cycle.
What’s the cooling challenge for tropical data centres?
The engineering story underpinning Southeast Asia’s data centre expansion deserves more attention than it gets. Traditional data centres operated at rack densities of 8–12 kW. AI workloads are pushing that to 25–40 kW and beyond (Turner & Townsend, 2026). Higher density means more heat. More heat in a tropical climate means more energy spent on cooling.
Data centres should ideally be maintained between 18°C and 27°C, according to the American Society of Heating, Refrigerating and Air-Conditioning Engineers. Southeast Asia’s year-round ambient temperature range of 27–35°C makes heat rejection significantly harder than in temperate climates. Humidity compounds the problem — it increases condensation and corrosion risk and reduces long-term equipment reliability.
The shift to liquid cooling has had the biggest impact across all major Southeast Asian markets. Liquid cooling adds between $1.80 and $2.40 per watt depending on architecture — rear-door, in-row, or direct-to-chip. High-density AI specifications (40+ kW per rack) typically increase total construction costs by 18–22% over legacy baselines (Turner & Townsend, 2026).
The cost equation for data centre construction in the region is shifting from civil works and labour toward power infrastructure, cooling systems, and imported long-lead equipment. Cost certainty now depends on procurement certainty — and procurement lead times for specialised cooling and power equipment are extending.
What does the energy gap mean for Southeast Asia’s AI ambitions?
Southeast Asia will account for a quarter of growth in global energy demand by 2035, according to the IEA. The data centre buildout is a major driver of that increase.
Several Southeast Asian nations are reviving nuclear power programmes specifically to address this gap. Malaysia set a 2031 target. The Philippines, building on earlier Bataan nuclear plant infrastructure, is actively pursuing atomic energy. Vietnam and Russia advanced a nuclear power cooperation deal in early 2026. Cambodia’s latest national strategy signalled openness to nuclear. Even Singapore and Brunei have indicated interest in studying nuclear potential.
Southeast Asia has never produced a single watt of nuclear energy. Whether these programmes can move from ambition to generation capacity on a timeline that matches data centre demand is the region’s most consequential infrastructure question. If the energy gap isn’t closed, the data centre boom hits a hard ceiling — regardless of how much capital is available for construction.
What’s the strategic picture?
The most interesting dynamic in Southeast Asia’s data centre expansion isn’t the construction itself — it’s the layering of US and Chinese technology infrastructure across the same set of countries. Malaysia hosts one of the largest concentrations of Chinese-owned data centre investments outside mainland China. Singapore is anchoring both Microsoft’s and Google’s regional operations. Thailand is receiving billions from ByteDance and the US hyperscalers simultaneously.
Tightening export controls on advanced semiconductors, growing scrutiny around technology supply chains, and questions about data governance and ownership are adding complexity to where AI-capable infrastructure can actually operate. The geography of data centre deployment increasingly shapes not just where enterprise workloads run, but which AI models are accessible, whose technology standards apply, and how data sovereignty is defined.
For companies operating across the region, this means a more fragmented infrastructure landscape where cloud platform availability, AI model access, and regulatory requirements vary meaningfully from market to market. That fragmentation is the cost of building fast in a contested space. It’s also, for now, the price of not falling behind.
Read the full AI Ecosystem Across Asia 2026 report — market data, country analysis, company profiles, investment landscape, and 2026–2030 outlook → digitalinasia.com/reports
