Image source: SIngtel.
Singtel has said that it plans to launch GPU-as-a-Service (GPUaaS) in Singapore and Southeast Asia by the third quarter of the year.
Powered by NVIDIA H100 Tensor Core GPU-powered clusters that are operated in existing Singtel data centres in Singapore, the telco will also be amongst the world's first to deploy NVIDIA’s next-generation GB200 Grace Blackwell Superchips, with Singtel amongst the first companies in the world set to receive them early next year.
The GPUaaS will then be expanded to run in the three new sustainable, hyper-connected, high-density AI data centres in Singapore, Thailand and Indonesia when they begin operations under Singtel’s regional data centre business, Nxera.
Bill Chang, CEO of Singtel’s Digital InfraCo unit and Nxera regional data centre business, said:
We are seeing keen interest from the private and public sectors which are raring to deploy AI at scale quickly and cost-effectively. Our GPUaaS will run in AI-ready data centres specifically tailored for intense compute environments with purpose-built liquid-cooling technologies for maximum efficiency and lowest PUE, giving them the flexibility to deploy AI without having to invest and manage expensive data centre infrastructure.
The Blackwell GPU is up to 30X faster in real-time large language model inference than their predecessors, 4X faster in training performance, and ans provides a 25X reduction in energy consumption and operating cost for LLM inferencing coupled with a new compiler.
Overall, it is hoped that Singtel’s GPU-as-a-Service offering, powered by NVIDIA, will help businesses across Southeast Asia to unlock the many efficiencies AI brings.
Our articles may contain affiliate links. If you buy through these links, we may earn a small commission.