(Nov 11): With artificial intelligence (AI) architecture becoming increasingly demanding, network operators and data centres must view supply chain planning as a key component of infrastructure design, says Simin Sirun, regional business director for Asean, India and Korea at electrical and network infrastructure solutions firm Panduit.
The overall supply chain in the Asia-Pacific region is growing more complex due to rising investments in AI, cloud infrastructure, advanced manufacturing and digitalisation, as well as data centres, which have necessitated specialised materials and manufacturing capacity.
In Malaysia, this is further amplified by the rate at which projects are being approved and getting underway. Additionally, research from Fitch Solutions company BMI estimates that the country has 4.6GW of planned data centre capacity under construction, with a tilt towards AI-centred projects.
“Customers increasingly require infrastructure to be delivered within compressed construction and energisation schedules. That places pressure not only on component availability, but also on forecasting, qualification, logistics and installation capacity,” says Sirun.
Sirun advises operators to integrate supply chain planning into infrastructure design to avoid unnecessary builds arising from fragmented planning. This can be done by defining thearchitecture early and using standardised, scalable connectivity platforms.
Doing so can ease one of the biggest challenges he sees in the physical supply chain, where failure can occur at any point across demand forecasting, production capacity, component qualification, ogistics and final installation, if specifications and alternatives cannot be aligned with the construction schedule due to a lack of early planning.
Localisation is another solution cited to secure supply chains. “The transition [to localisation] should begin with early supplier engagement, standardised designs, qualified alternative configurations and stronger collaboration among developers, consultants, contractors and infrastructure providers.
“Localisation should ultimately improve responsiveness and resilience without compromising quality, interoperability or long-term performance,” he says.
Sirun also addresses the supply chain needed for the planned 4.6GW of data centre capacity, saying that the main structural risk lies in the potential misalignment between the pace of data centre development and the readiness of the broader supporting ecosystem, meaning its ability to support scaling in a coordinated and reliable manner.
Though power availability is rightly considered an important concern for data centres, operators must not neglect transmission and distribution capacity, planning approvals, skilled resources, connectivity, cooling and the availability of critical infrastructure components if they are to ensure a successful data centre deployment.
Sirun says Malaysia should prioritise coordinated planning among government agencies, utilities, developers, technology providers and the supply chain for the smooth deployment of data centres.
He adds that the country has strong fundamentals in its manufacturing ecosystem, skilled talent and strategic connectivity to regional markets.
“Malaysia’s long-term potential and opportunity is significant, but disciplined infrastructure planning will be essential to convert investment commitments into reliable, operational capacity.”
For Malaysian data centre operators to safeguard their target energisation timelines, Sirun says the most important shift will be moving from transactional procurement to lifecycle- and schedule-based procurement.
“In an AI data centre project, selecting infrastructure only on the basis of unit price can create greater risk if the product cannot be supplied, integrated, installed or expanded within the equired timeframe,” he adds.