Beyond the Billion-Dollar Announcements: Is Malaysia Building AI or Just Housing It?

Malaysia’s AI infrastructure push is accelerating — but questions remain over whether the country can translate foreign-led data centre growth into genuine domestic AI capability and technological sovereignty. Photo/graphic credit: Graphic: TNS News Analysis

Equinix’s new KL2 data centre in Cyberjaya is the latest — and largest — signal that global operators are doubling down on Malaysia. But as investment pours in, a sharper question is emerging: who captures the real value?

By TENGKU NOOR SHAMSIAH TENGKU ABDULLAH

KUALA LUMPUR, May 17 – When Equinix announced last Monday that it would build a fourth data centre in Malaysia, a US$190 million facility in Cyberjaya to be known as KL2, the news was quickly filed away as another entry in the country’s expanding ledger of foreign digital infrastructure commitments. The headlines wrote themselves: fourth data centre, 2,200 cabinets, AI-ready, liquid cooling, renewable energy from day one.

But look at the KL2 announcement in the context of what has been happening to Malaysia’s data centre sector over the past 18 months and it starts to look less like a routine expansion and more like a deliberate strategic signal, one worth reading carefully.

Timing the bet

The land for KL2 was quietly acquired back in July 2024, spanning 14,300 square metres a parcel Equinix secured for around RM23 million. The public announcement came only this week. That gap between land acquisition and public commitment is telling. It means Equinix was watching, and what it saw evidently justified proceeding.

What was it watching? A market in the middle of a decisive policy pivot. Prime Minister Datuk Seri Anwar Ibrahim confirmed earlier this year that Malaysia had stopped approving non-AI-related data centre proposals since 2024 -z a clear signal that the country is moving towards larger, more capital-intensive projects aligned with national AI ambitions.

For a company like Equinix, whose business model is built on colocation and interconnection rather than hyperscale AI factories, that policy environment creates both opportunity and risk. KL2 with its liquid cooling capacity for high-density AI and HPC workloads appears calibrated precisely to qualify under the new dispensation.

The cooling question

Liquid cooling is more than a technical feature it is increasingly the dividing line between data centres that can support serious AI workloads and those that cannot. But deploying the technology is one thing; sustaining it is another, and this is where Malaysia faces a challenge that rarely features in the press releases.

An academic specialising in AI and electrical engineering at a Malaysian university, who requested anonymity due to institutional constraints, flags this gap directly. “Liquid cooling is a critical foundation for supporting advanced AI workloads, particularly in Malaysia’s hot and humid climate,” the academic told TNS News.

“However, several challenges persist, including a shortage of local expertise in advanced thermal and fluid system engineering, the complexity of operating such systems such as managing flow balance and preventing leaks high upfront investment and retrofit difficulties, as well as broader constraints related to water resources and power infrastructure.”

That last point deserves particular attention. Malaysia’s tropical climate makes effective cooling both more necessary and more technically demanding than in temperate markets. The engineering gap the academic identifies is not a reason to slow investment but it is a reason to invest in parallel in the talent pipeline that will ultimately determine whether this infrastructure performs as intended.

The interconnection play

What distinguishes Equinix’s approach from pure hyperscale competitors is the interconnection layer. KL2 will not sit in isolation, it will link Equinix’s Malaysian data centres with its Singapore campus via Equinix Fabric, enabling enterprises to deploy high-performance AI workloads and distributed data strategies across ASEAN.

This matters because the Kuala Lumpur–Johor–Singapore corridor is increasingly being treated as a single infrastructure zone rather than three discrete markets. More than two-thirds of Equinix’s existing Malaysian customers already operate across multiple metro locations, a pattern that reflects how regional businesses actually consume digital infrastructure today.

Malaysia’s proximity to Singapore has long been a structural advantage. Singapore’s own restrictions on new data centre developments after a three-year moratorium, the city-state now enforces rigid green data centre specifications redirected significant investment momentum southward, helping Malaysia emerge as the region’s fastest-growing data centre market.

Equinix is positioning KL2 to capture not just Malaysian demand but the overflow and extension of Singapore-anchored ecosystems.

The harder questions

None of this unfolds against a frictionless backdrop. Grid connection delays not electricity generation capacity are the primary bottleneck for new data centre developments in Malaysia, with some electricity and water approvals taking up to 18 months. Johor was said to have rejected around 30 percent of new data centre applications in 2024 due to misaligned utility timelines and planning complications.

There is also a utilisation question that the government has not yet resolved. Parliamentary data from November 2025 showed that Malaysia’s data centres were consuming only 603 megawatts against a declared maximum demand of 1,276 megawatts roughly 47 percent against a government target of 85 percent utilisation. The gap between announced capacity and actual consumption has prompted concerns about speculative development and stranded assets.

And then there is the deeper structural question one the same academic frames with a precision that no amount of corporate messaging can paper over. “The infrastructure strategy is necessary but not sufficient,” the academic says. “Without concurrent investment in talent, research, and local AI firms, Malaysia risks scaling capacity without capturing commensurate economic and technological value.”

It is a warning that cuts to the heart of the sovereignty debate. A January 2026 analysis warned that the bulk of Malaysia’s engagement with AI development runs through foreign investment from multinational hyperscalers, raising the risk that the country ends up subsidising the computational needs of foreign corporations without securing commensurate returns in local innovation capacity or sovereign technological control. The academic’s formulation necessary but not sufficient — may be the most useful lens through which to read not just the KL2 announcement, but every billion-dollar commitment that has preceded it.

The market backdrop

Asia Pacific’s data centre capacity is forecast to expand from 11GW in 2025 to 15GW by 2028, growing at an average annual rate of 11 percent, fuelled by accelerating AI adoption and cloud expansion across major economies. Within that regional picture, Malaysia’s data centre market valued at USD 6.14 billion in 2025 is projected to more than double to USD 13.57 billion by 2030.

Malaysia’s pipeline now consists almost entirely of AI-focused projects, with approximately 4.6GW of planned and under-construction capacity, a concentration that has raised barriers to entry and strengthened the dominance of well-funded global operators. KL2, with its AI-ready specifications and renewable energy commitments from day one, fits squarely within that profile.

What it adds up to

The KL2 announcement is, at one level, a straightforward investment in growing demand. At another level, it is a statement about where Malaysia’s digital infrastructure story is heading interconnected, AI-oriented, sustainability-bound, and increasingly integrated into the broader ASEAN corridor.

Whether Malaysia successfully converts that infrastructure wave into genuine technological sovereignty or remains, as critics caution, primarily a host economy for foreign computational needs is the larger question no single data centre announcement can settle.

But the academic’s assessment offers the clearest formulation of what is actually at stake: the infrastructure is necessary. Without the talent, the research, and a thriving ecosystem of local AI firms growing alongside it, however, the capacity Malaysia is racing to build may ultimately serve someone else’s future more than its own.

  • TNS NEWS

Tengku Noor Shamsiah Tengku Abdullah is Editor-in-Chief of TNS News and Founder of TNS Consulting, a strategic business and development consultancy operating under Kuala Lumpur-based TNS Media Consultant. She has covered Malaysian and regional affairs for more than 30 years and was previously based in Singapore.

Leave a Reply

Discover more from TNS News

Subscribe now to keep reading and get access to the full archive.

Continue reading