Malaysia emerges as key AI hub amid data centre boom

While Silicon Valley debates which AI model will reign supreme, Malaysia is quietly winning the race for where all that intelligence actually gets computed. The country has approved roughly RM184.7 billion, about $42 billion, in data center investments between 2021 and 2024, transforming itself from a peripheral player in Southeast Asian tech into one of the region’s most critical infrastructure hubs.

The Malaysian data center market was valued at approximately $4.04 billion in 2024. By 2030, that figure is projected to exceed $13.5 billion. The AI-optimized segment alone is forecast to grow from $0.49 billion in 2025 to $1.76 billion by 2030, a compound annual growth rate of 29.03%.

The Johor corridor

The epicenter of this boom sits in Johor, Malaysia’s southernmost state, which shares a border with Singapore. Johor alone had 42 data center projects worth RM164.45 billion approved by the second quarter of 2025, making it the leading state for data center growth across all of Southeast Asia.

The state now boasts over 1,600 MW of installed IT capacity. Singapore has the financial infrastructure, the talent pipelines, and the connectivity that hyperscalers crave, but it’s running out of room and cheap power. Johor offers land, lower operational costs, and sits close enough to Singapore that fiber latency between the two is negligible.

Big tech goes all in

Google has committed $2 billion to data center projects in Malaysia. Microsoft has pledged $2.2 billion. And in a deal that underscores just how serious the AI infrastructure play has become, YTL Power partnered with NVIDIA on a $2.36 billion AI infrastructure project in Kulai, Johor, with completion targeted for October 2025.

National data center capacity across Malaysia is expected to grow from 0.9 to 1 GW in 2025 to somewhere between 3 and 4 GW by 2029.

The energy problem nobody wants to talk about

Beginning in February 2026, Malaysia will impose restrictions on new non-AI data center projects. Traditional colocation and cloud facilities that don’t serve AI workloads will face a higher bar for approval, while AI-focused projects get priority access to power and permitting.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our Editorial Policy.