- TM Global said current data center capacity is inadequate to meet the growing demands of AI
- From fiber upgrades to cross-border connectivity and its PWR 2030 strategy, TM is deeply integrating AI
- TM Global aims to position Malaysia as a regional digital hub
Turns out the data center capacity shortage is a global phenomenon. TM Global, part of Telekom Malaysia, which launched GPU-as-a-service (GPUaaS) in December, told Fierce demand for GPU-driven AI applications is outpacing the country’s current data center capacity.
“Power-hungry GPUs require data centers with sufficient capacity, particularly Tier 3 or Tier 4 facilities," TM Global VP Baharul Nizam Said Daliman said. "While the industry is building data centers, it is not enough to fulfill the market demand."
Malaysia has a vibrant data center industry with several players — including Microsoft Azure, Google Cloud and Amazon Web Services — and growing competition. TM Global is just one of the players expanding its footprint in the country to meet rising order levels.
But TM Global said there's something that sets it apart from its hyperscale competitors: its telco roots. The ability to offer diverse, scalable routes to its data centers gives it an edge in a competitive market. Sure, this can be done by others, but not at the scale TM Global can do it, Nizam claimed.
“In Malaysia, most of the data center builders are not telcos and telcos are not data center builders. We wanted to combine the two aspects, which is a unique proposition in Malaysia," Nizam said.
TM Global's expansion plans
The company plans to develop multiple data center clusters in Johor, a state in southern Malaysia, and a few in Klang Valley. As part of that effort, it's also beefing up its network.
“We are modernizing our core and fiber network to take the infrastructure in the greenfield areas as well, because some data centers are being developed outside the main industrial areas based on the availability of land, water or power,” Nizam explained. The company is also planning to introduce more high-end services, such as managed wavelength for data transmission between data centers or clusters, he added.
Strengthening its network will be key in delivering the low-latency, high-capacity connectivity required by AI applications. That network also includes subsea and international connections.
“We continue to invest in the international submarine and also cross-border connectivity, with Singapore, for example. Apart from Singapore, we are working with our partners in Thailand to connect data centers from Bangkok, Malaysia and Singapore,” Nizam said.
The company has collaborated with Singapore’s largest service provider, Singtel’s data center arm, Nxera, to develop data centers in Malaysia, starting with an AI data center campus in Johor. The initial planned capacity of this data center is 64MW with plans to scale up to 200MW.
To be clear, TM Global is not planning to build data centers in other countries in the region but “working with Singtel and Nxera enables us to extend our footprint, if required, since they have presence in Thailand, Batam and Jakarta,” among others, Nizam said.
Apart from its partnership with Nxera, the company believes it can play a bigger role in the Association of Southeast Asian Nations (ASEAN), which comprises 10 Southeast Asian countries: Brunei, Cambodia, Indonesia, Laos, Malaysia, Myanmar, Singapore, Thailand, the Philippines and Vietnam.
“The ASEAN region has a population of around 400 million, and at TM we have infrastructure that goes in most of this region. We have POPs [points of presence] in Bangkok in Thailand, Singapore and Jakarta in Indonesia. We have also invested in a submarine cable connecting Brunei, the Philippines and Vietnam. Our plan is to position Malaysia as a regional digital hub that supports the growing demand for AI,” Nizam said.
TM Global's AI strategy
Zooming out, parent company Telekom Malaysia is looking to leverage AI to improve its own operations. Its PWR 2030 strategy, launched in 2024, is built on three core principles (power, win and revitalize) aimed at transforming the company into a digital powerhouse by 2030. AI is a significant part of this strategy.
The company’s capital expenditure was $380 million (RM1.6 billion) last year, which was spent on “expanding access, enhancing network efficiency and fortifying our digital capabilities — driving both TM and the nation’s growth.”
“There are multiple ongoing AI initiatives,” Nizam said. “All divisions will have their own AI use cases and in some, we will use agentic AI going forward. Rather than reducing the workforce, we are focusing on upskilling and growing the capabilities of our workforce.”