Singtel, a leading global communications technology firm, is developing advanced data centers to address the growing need for AI workloads. In response to the exponential growth of AI use across multiple industries, the company is creating sustainable, hyper-connected, high-power density GPU management facilities.
Bill Chang, CEO of Singtel's Digital InfraCo unit said, “With the fast-growing adoption of AI across various industries, we see an inflection point in the data center industry for high power density, yet sustainable and hyper-connected AI DCs that are optimized to manage GPU clusters for AI workloads. This is why we are building our fourth generation data center platform specifically tailored for such intense compute environments.”
Singtel’s Current and Future Developments
The first fourth-generation AI data center, DC Tuas, is under development with 58MW of capacity. The facility boasts great power density and innovative sustainability elements like liquid cooling and lower-carbon incorporated construction materials. With a full-load PUE (Power Usage Effectiveness) of 1.23, DC Tuas strives for energy efficiency. Singtel is also switching its data centers to green and renewable energy to reduce its carbon footprint.
Singtel is also developing two more state-of-the-art data centers in Thailand and Indonesia which will increase its total pipeline capacity to over 200MW; adding to its 62MW in Singapore. The company received a SGD 535 million five-year green loan to expand operational capacity.
Singtel's new generation AI DCs will use direct-to-chip liquid and immersion cooling to efficiently manage GPU cluster heat. To manage expensive GPU clusters over various networks, real-time server monitoring platforms are essential. The Paragon Platform™ will connect these AI data centers to Singtel's fixed and 5G mobile networks, enabling high bandwidth and low latencies for diverse industry use cases.
Mr Chang added, “Within these AI DCs, we will offer a host of core services including cloud co-location services, managed hosting services and value-added GPU-as-a-Service (GPUaaS). We see growing demand from both enterprises and cloud companies to leverage our GPU cluster to complement their environments with greater agility in these AI DCs as well.”
Singtel has developed and implemented GPUaaS in the 5G@Sentosa testbed with GovTech and other government organizations since 2021. GPUaaS and Singtel's leadership in data centers, fixed and 5G networks allow enterprises to accelerate the development of generative AI, large language models, and other AI workloads without having to invest in AI DC space and GPUs. AI helps businesses and government agencies concentrate and accelerate innovation and transformation.
Recognizing the widespread adoption of AI across various industries, Singtel is strategically crafting advanced facilities designed to be sustainable, hyper-connected, and equipped with high-power density GPU management capabilities. These developments signify Singtel's commitment to providing state-of-the-art infrastructure that not only supports the burgeoning requirements of AI but also aligns with sustainable practices, fostering efficiency and connectivity in the rapidly evolving digital landscape.