CIMB Investment Bank reports that Telekom Malaysia Bhd (TM) is considering plans to construct a new hyperscale data centre (HDC) in Malaysia.
This initiative aims to enhance TM’s capacity beyond its current Klang Valley Data Centre (KVDC) and Iskandar Puteri Data Centre (IPDC), according to The Edge Malaysia.
By utilising its extensive submarine and terrestrial networks along with established local partnerships, CIMB suggests that TM could attract robust international partners to co-invest in the HDC.
Such partnerships could not only provide financial support but also contribute expertise in designing and operating data centres tailored for hyperscalers and artificial intelligence (AI) workloads.
CIMB highlights potential enhancements such as more efficient cooling solutions like direct-to-chip liquid or immersive cooling, as well as advanced real-time server monitoring platforms to manage costly graphics-processing unit clusters and prevent overheating.
Additionally, CIMB speculates that strategic partners could assist in attracting tenants for the HDC by offering expansion space to their current data centre tenants.
If TM proceeds with the project, CIMB anticipates a phased development approach, with the initial phase possibly commencing in the second half of 2024 and becoming operational by the first half of 2026.
What is HDC?
According to Nlyte Software, an HDC serves as a facility that accommodates essential compute and network infrastructure. It plays a crucial role in enabling companies such as Amazon, Google, and Microsoft to harness substantial processing power, facilitating the delivery of vital services to customers on a global scale.
Hyperscale facilities are generally characterised by having a substantial capacity, often comprising a minimum of 5,000 servers and encompassing a physical space of 10,000 square feet (~929 square meters) or more.
These expansive data centres are designed to meet the growing demands of large-scale cloud service providers and internet companies, ensuring they can efficiently handle the immense computational requirements of delivering services worldwide.
