As artificial intelligence moves from pilot projects into enterprise operations, the question facing businesses is no longer whether to adopt AI, but how to build the infrastructure that makes it work reliably, efficiently and at scale. That question was at the centre of the discussion when Mr Chan Kee Siak, Founder & CEO of Exabytes Group, took the stage at KL20@Penang as a featured speaker in the Future Tech Conversations series.
KL20@Penang: A Conversation Built for Technologists
Under Track 03, CONNECT | Exchange Conversations, the Future Tech Conversations series brought global and industry technology experts together to look at the breakthroughs reshaping semiconductors and AI. The topics ranged from advanced packaging and intelligent manufacturing to compute architectures, enterprise deployment and next-generation infrastructure. The sessions were designed for technologists, engineers, technical leaders and R&D professionals, and focused on the engineering challenges and trade-offs that will define the next era of semiconductor and AI computing.
Mr Chan spoke in the session “Engineering the Infrastructure for AI at Scale,” moderated by Ts. Jaffri Ibrahim, CEO of CREST. He was joined by Thum Siew Beng, Senior Director and Malaysia Engineering Site Lead at Microsoft. Together, they looked at the full infrastructure stack needed to train, deploy and run AI workloads: accelerated computing, high-performance networking, cloud architecture, workload orchestration and enterprise deployment.
AI Performance Is a Systems Problem
The conversation started from a simple point: as AI workloads grow in scale and complexity, performance depends more and more on how compute, networking, storage and cloud infrastructure are engineered to work as one system.Drawing on more than two decades of building digital infrastructure for businesses across the region, Mr Chan offered a practical view. Powerful GPUs matter, but they are only one part of the equation. Bottlenecks often appear where the layers meet. Data pipelines can’t feed compute fast enough. Networks struggle with east-west traffic. Storage architectures weren’t designed for AI-scale throughput. Orchestration layers can’t allocate resources efficiently across mixed workloads.
“The conversation around AI often starts with GPUs, but performance is decided at the seams: how fast data moves, how workloads are orchestrated and how securely everything runs together. Getting those connections right is what turns AI investment into business results.” — Chan Kee Siak, Founder & CEO, Exabytes Group
From Hyperscale to Main Street: Making AI Accessible
Thum Siew Beng brought a hyperscaler’s engineering perspective. Mr Chan added a view that matters just as much for Malaysia’s digital economy: how AI infrastructure reaches the businesses that make up most of the economy.
Small and medium enterprises make up the vast majority of businesses in Malaysia and Southeast Asia. For them, the challenge is less about building data centres and more about getting reliable, affordable and secure access to AI capabilities that fit into how they already work. This is where the digital solutions layer matters: cloud hosting, managed infrastructure, cybersecurity and business applications, delivered in a way that removes complexity instead of adding it.
This thinking sits behind Exabytes’ GROW AI direction, which aims to help businesses move from AI curiosity to AI capability. The focus is on practical adoption, on infrastructure that scales as businesses grow, and on solutions that deliver measurable outcomes rather than experimental ones.
“Our role is to take the complexity out of AI. Hyperscalers and chipmakers are pushing the frontier. Exabytes’ job is to make sure that progress reaches the businesses that need it most, in a form they can adopt, afford and trust. That’s what GROW AI is about.” — Chan Kee Siak, Founder & CEO, Exabytes Group
Where the Layers Must Work Together
A key theme of the session was the need for collaboration across the ecosystem. No single layer, and no single organisation, can deliver AI at scale alone. Semiconductor innovation, hyperscale cloud platforms, regional infrastructure providers, systems integrators and enterprises all have to work in concert.
For Malaysia, this is a real opportunity. Penang has long been a global hub for semiconductor manufacturing and engineering talent. Hosting this conversation there highlighted how the country can connect upstream hardware innovation with downstream digital adoption. Local digital infrastructure players like Exabytes help close that loop by turning advanced capabilities into services that businesses can actually use.
Looking Ahead
Mr Chan’s participation at KL20@Penang reflects Exabytes’ ongoing commitment to shaping the conversation on AI and digital transformation in the region. Exabytes serves more than 160,000 customers worldwide. As businesses face growing pressure to adopt AI responsibly and competitively, Exabytes continues to invest in the infrastructure, solutions and partnerships that make that transition possible.
The message from the stage was clear. The future of AI will be decided not only by breakthroughs in chips and models, but by how well the whole infrastructure stack is engineered, integrated and made accessible. Through GROW AI and its broader portfolio of digital solutions, Exabytes intends to help businesses across Southeast Asia be ready for that future.
To learn more about how Exabytes can support your AI and digital transformation journey, visit www.exabytes.ai
