In an interview with RCR Wireless News, Hafiz Lockman, general manager of international sales at TM Global, described the diverse range of enterprise and public sector clients already adopting its GPUaaS offering
In sum – what you need to know:
Sovereign GPUaaS demand surges – TM Global’s sovereign GPU-as-a-Service is seeing strong uptake in sectors like healthcare, media, and logistics, driven by regulatory and real-time AI needs.
Regional edge & fiber expansion – TM is addressing cross-border AI challenges by growing its edge data center footprint and enhancing fiber routes in the ASEAN region.
Security-first, modular design – Businesses want customizable, secure infrastructure—TM aims to meet that with flexible GPU options, high-speed fiber, and compliance-ready Malaysian data centers.
TM Global, the wholesale division of Malaysian operator Telekom Malaysia (TM), is seeing rapid uptake of its sovereign GPU-as-a-Service (GPUaaS) platform, which delivers on-demand access to high-performance computing infrastructure for real-time AI workloads.
The recently launched service seems to be gaining traction across sectors including healthcare, public safety, immersive media, manufacturing, and logistics—industries where low-latency and data sovereignty are paramount.
In an interview with RCR Wireless News, Hafiz Lockman, general manager of international sales at TM Global, described the diverse range of enterprise and public sector clients adopting the service.
“At TM Global, we are seeing immediate demand for GPU-as-a-Service (GPUaaS) across key sectors where real-time analytics and high-performance computing are mission-critical,” Lockman said. “These include healthcare (for diagnostics and predictive analytics), immersive media (for content rendering and real-time processing), public safety and smart city services (for surveillance, traffic optimization, and predictive modeling), and autonomous systems in manufacturing and logistics.”
The GPUaaS platform is fully hosted in Malaysia, leveraging TM’s AI-ready, LEED-certified data centers and backed by its nationwide high-speed fiber network. The infrastructure allows companies to access scalable GPU power without heavy upfront investments in on-premises computing, the executive said.
According to Lockman, a core driver of adoption is the platform’s flexibility. Businesses can customise their environments by selecting a tailored mix of data center space, hardware, GPU chips, and connectivity.
“Early adopters of GPUaaS have consistently highlighted the need not only for greater processing power, but also for enhanced flexibility and scalability to keep pace with the fast-evolving demands of AI workloads,” said Lockman.
“While they value the ability to access high-performance Nvidia GPUs on demand, many have expressed a clear preference for modular, customisable solutions that allow them to mix and match only what they need,” he said.
As AI and edge computing expand across ASEAN, connectivity remains a critical challenge. Many enterprises face inconsistent network performance and complex cross-border regulatory environments.
Lockman explained how TM is tackling these hurdles: “From TM Global’s engagements across the region, key challenges include fragmented infrastructure, variable network performance, and regulatory diversity. These issues often result in high latency, unpredictable service levels, and increased complexity in deploying AI or edge applications at scale across borders.”
TM is expanding its regional footprint through upgrades to IPDC and KVDC facilities, and by deploying a growing number of edge data centers throughout Malaysia. It has also extended its global edge node presence to around 160, enabling low-latency processing for real-time AI, live streaming, and cloud gaming applications across Southeast Asia.
Additionally, the company is strengthening fiber routes along the vital Singapore–Malaysia–Thailand–Indonesia corridor and taking advantage of its operator license in Singapore to build direct interconnections with regional carriers and hyperscalers, the executive added.
One major step forward is the Nxera–TM data center joint venture in Johor Bahru, slated to go live by 2026. The AI-optimized campus will start with 64 MW and can scale to 200 MW, offering high-density GPU infrastructure and regional AI capacity for hyperscalers.
“Hosted entirely within Malaysia’s borders, [out GPUaaS solution] aligns with data residency laws and regulatory standards, ensuring that data remains under local jurisdiction,” Lockman said.
The service is backed by secure, energy-efficient data centres designed for continuous monitoring, proactive threat mitigation, and disaster recovery readiness, he said.
“Data security is a top priority for businesses, and TM’s data centers are purpose-built with this in mind. Adopting a ‘security by design’ approach, our facilities offer continuous protection through real-time monitoring, automated threat mitigation tools, and proactive cyber risk management,” the executive added.
