The Asia-Pacific region's data center capacity is set to expand rapidly, growing at 11% annually from 11GW to 15GW between 2025 and 2028, fueled by exploding AI demands.
Southeast Asia, particularly Malaysia, is emerging as a hotspot with a massive 4.6GW pipeline of planned and under-construction facilities, most tied to AI workloads.
Malaysia Emerges as AI Data Hub
Johor state in Malaysia already boasts around 500MW of live capacity, drawing huge investments from tech giants like Microsoft and Google, each pledging over $2 billion.
Since 2024, Malaysia has halted approvals for non-AI data centers and rejected about 30% of proposals due to unsustainable power and water demands.
New policies like the Data Center Framework and "85% utilization rule" prioritize high-value AI projects with strict sustainability mandates.
Singapore Tightens Efficiency Standards
Singapore, with 1.5GW operational capacity, launched a framework allocating at least 200MW to operators achieving low power use and high green energy reliance.
Regional Investments Accelerate
In Indonesia, Microsoft's $1.7 billion commitment supports over 14 cloud zones, while Thailand eyes ByteDance's $4 billion push.
Major deals include Blackstone's $16 billion acquisition of AirTrunk and Singtel/KKR's $6.6 billion stake in ST Telemedia Global Data Centres.
Operators like Bridge Data Centres innovate with sustainable fuels, piloting HVO-powered backups to meet green goals.
Challenges and Broader Impacts
This boom strains power grids, prompting pilots like Thailand's 2GW direct power purchase and sparking nuclear power discussions across the region.
For everyday people, it means more jobs in construction and tech, faster AI services like chatbots and personalized apps, but also rising energy costs and urban changes near sites like Johor.
Looking ahead, mergers and modular builds will dominate, but success hinges on balancing growth with renewable energy to avoid blackouts and environmental backlash.
Geopolitically, Southeast Asia's rise diversifies supply chains away from tense hotspots, securing global AI infrastructure for the future.
Andrew Lee
