For years, Singapore served as the undisputed core of Southeast Asia’s digital infrastructure. However, the immense power and space requirements of modern artificial intelligence workloads are forcing a regional rebalancing. As mature hubs grapple with strict grid constraints and limited land, Malaysia has emerged as a compelling alternative. Driven by a combination of scalable power, vast land availability, and strategic geographic proximity, states like Johor are experiencing an unprecedented boom in hyperscale cloud and colocation development. According to recent reports, billions of dollars in foreign direct investment are flowing into the country to build next-generation facilities designed specifically for the intense demands of AI.
Yet, erecting vast physical structures and securing grid connections is only half the battle. Inside these hyperscale facilities, the rise of AI introduces a new layer of complexity in managing the enormous volumes of east-west traffic generated between servers and computing clusters. The underlying physical network infrastructure, from high-density cabling to precision optical glass, has become a vital differentiator in preventing costly computational bottlenecks. Supply chain resilience, rapid deployment timelines, and the ability to maintain long-term performance standards are now defining the success of these massive digital investments.
Read our thoughts on Malaysia’s AI and data centre play. Can policy keep up with demand?
To better understand this physical shift, we speak to Kevin Choong, Senior Business Manager for Network Infrastructure Unit, Malaysia, Indonesia and Singapore, Panduit, about the foundational role of network connectivity in the AI era. In this interview, we aim to uncover how the industry is managing raw material shortages, why passive physical infrastructure is critical to preventing compute bottlenecks, and how the company is upskilling local talent across Southeast Asia to keep pace with these highly complex deployments.
What key infrastructure advantages are driving the rapid shift of data centre investments from traditional hubs like Singapore to Malaysia?
Malaysia offers a compelling combination of land availability, competitive power costs, and the ability to support large-scale developments that are increasingly difficult to accommodate in mature markets such as Singapore.
Over the past few years, hyperscale cloud providers, AI operators, and colocation companies have required significantly larger campuses with access to scalable power. Johor, in particular, benefits from its proximity to Singapore while providing the space and resources needed for next-generation facilities.
Additionally, Malaysia continues to attract investment through supportive government initiatives, improving digital infrastructure, and strong connectivity to regional and global networks. As a result, we see Malaysia evolving into one of Southeast Asia’s key digital infrastructure hubs.
With data centres generating over 50% of your revenue in Malaysia, is this hyper-growth sustainable over the next 3 to 5 years given power grid constraints?
While power availability is an important consideration for every data centre project, it is ultimately a matter managed by developers, operators, utilities, and government stakeholders.
From Panduit’s perspective, what we continue to see is strong demand for high-performance network infrastructure driven by cloud adoption, AI workloads, enterprise digital transformation, and increasing data consumption across the region.
Regardless of where individual projects are developed, customers are placing greater emphasis on scalable fibre infrastructure, higher-density connectivity, and future-ready physical networks. Those trends are expected to continue over the coming years.
Our focus is helping customers deploy reliable and scalable infrastructure that can support both current requirements and future growth, while working closely with the broader ecosystem to ensure projects can be delivered successfully.
With enterprise clients prioritising fast delivery over price, how are you managing raw material shortages, like optical glass, to maintain short lead times?
Supply chain resilience has become a strategic differentiator in the AI era. At Panduit, we work closely with our global manufacturing and supplier network to ensure visibility across demand forecasts, inventory planning, and production capacity. We also maintain strategic relationships with key material suppliers and continuously diversify sourcing where appropriate.
Rather than reacting to shortages, we focus on proactive planning with customers and partners to align project schedules early. This approach helps us maintain reliable lead times even during periods of elevated demand.
Ultimately, our customers value certainty and speed of deployment, and that is where robust supply chain management becomes critical.
What critical role does passive physical infrastructure, like optical glass and cabling, play in preventing bottlenecks in AI workloads?
AI performance is often discussed in terms of GPUs, processors, and software, but the physical network infrastructure is equally critical.
AI clusters generate enormous volumes of east-west traffic between servers, storage systems, and networking equipment. If the underlying fibre infrastructure cannot support those data flows efficiently, expensive computing resources can become underutilised.
High-performance cabling systems help ensure low-latency communication, signal integrity, and scalability across the entire environment. In many cases, the success of an AI deployment depends as much on the quality of the connectivity architecture as it does on the computing hardware itself.
That is why we often describe physical infrastructure as the foundation that enables AI systems to achieve their full potential.
How are high-density components, such as SN connectors, being re-engineered to guarantee transmission performance across GPU-to-storage networks?
As network speeds continue to increase from 400G toward 800G and beyond, connectivity solutions must evolve to support higher density without compromising performance.
Newer connector technologies such as SN connectors are designed to optimise space utilisation while maintaining stringent optical performance requirements. The focus is on minimising insertion loss, improving reliability, and enabling efficient cable management in increasingly dense environments.
We are also seeing greater emphasis on interoperability and modular designs that allow operators to scale infrastructure rapidly while maintaining serviceability.
The objective is straightforward: ensure the physical layer can support the performance requirements of next-generation AI fabrics and storage architectures.
How do you balance hyper-accelerated deployment timelines (as fast as 11 months) with maintaining strict Tier certifications and 25-year performance warranties?
Today’s data centre projects are being delivered at unprecedented speed, particularly with the rapid growth of cloud and AI infrastructure. While Tier certifications are governed by the relevant design, construction, and certification stakeholders, the physical infrastructure deployed within the facility must consistently meet the required performance and reliability standards.
At Panduit, our role is to provide validated infrastructure solutions that support efficient deployment while maintaining quality and long-term reliability. This includes high-performance cabling systems, connectivity solutions, and best-practice installation methodologies developed through industry standards and extensive testing.
The key to successful accelerated deployments is early collaboration among developers, consultants, contractors, technology partners, and infrastructure providers. When all stakeholders are aligned from the design stage, projects can move quickly without compromising operational requirements or future scalability.
Ultimately, speed is important, but predictability and long-term performance remain equally critical for data centre operators.
How are you upskilling regional installation partners across Southeast Asia to ensure local technical capabilities keep pace with complex AI infrastructure requirements?
Technology innovation is moving extremely quickly, and workforce development must keep pace. Across Southeast Asia, we invest in ongoing training, certification programmes, technical workshops, and hands-on engagement with channel partners and system integrators. The focus is not only on product knowledge but also on emerging architectures supporting AI, high-density fibre deployments, at hyperscale data centre environments.
We believe local expertise is essential because successful AI infrastructure deployment depends on both world-class technology and the availability of skilled professionals who can design, install, and maintain these increasingly sophisticated environments.
Developing regional talent helps ensure customers receive consistent levels of quality, regardless of where projects are deployed across Southeast Asia.