Singtel and Nscale unlock GPU capacity globally

This collaboration aims to meet the growing global demand from enterprises for generative AI, high-performance computing and data-intensive workloads.

Singtel will leverage Nscale’s AMD and NVIDIA GPU capacity in Europe for Singtel’s customer workloads across key markets in the region. This capability ensures that Singtel can deliver to high-volume requirements on demand and maintain service excellence especially when additional capacity is needed. Correspondingly, Nscale will be able to tap into Singtel’s NVIDIA H100 Tensor Core GPU capacity in the Southeast Asian region for their customers’ workloads through an integration with Singtel’s patented orchestration platform, Paragon. Furthermore, as Singtel’s regional data centre arm Nxera expands in the region, its sustainable AI-ready data centres will provide the necessary data centre capacity to support large-scale deployment of Nscale GPU capacity.

This partnership will allow Singtel and Nscale to build out a more comprehensive GPU-as-a-Service (GPUaaS) offering globally, ensuring that their customers benefit from the flexibility of a wider geographic footprint and robust infrastructure support. This will also drive greater utilisation in their respective GPU clusters.

Mr Bill Chang, CEO of Singtel’s Digital InfraCo and Nxera, said: “As we continue to augment our GPUaaS offerings, we are forging a series of strategic partnerships to grow our ecosystem and broaden our service availability for our customers. Our partnership with Nscale will allow our customers to tap into their high-performance GPU resources on demand, unlocking new possibilities for innovation and efficiency. Our commitment to delivering cost-effective solutions, backed by our state-of-the-art data centres, ensures businesses can access high-performance GPU resources quickly and seamlessly.”

Mr Josh Payne, Nscale Founder and CEO said: “Nscale is the vertically integrated GPU Cloud building the global infrastructure backbone for generative AI. Our sustainable AI-ready data centre together with our GW pipeline of data centre capacity uniquely positions us to deliver sustainable AI infrastructure at any scale for customers worldwide. Through this strategic partnership, Nscale will provide Singtel customers with unmatched access to sustainable, high-performance, and cost-effective AI compute to accelerate enterprise generative AI in the region and beyond.”

Singtel previously announced in February that it will be launching its GPUaaS later this year, providing enterprises with access to NVIDIA’s AI computing power. This will enable them to deploy AI at scale quickly and cost effectively to accelerate growth and innovation. Singtel also recently announced a partnership with Vultr in the US and a strategic partnership with Bridge Alliance that will bring its GPUaaS offerings to enterprises across Southeast Asia.

Singtel’s GPUaaS will be expanded to run in new sustainable, hyper-connected, AI-ready data centres by Nxera across Singapore, Thailand, Indonesia and Malaysia when they begin operations from mid-2025 onwards.

Nscale's strategic partnership with Singtel follows a number of recent announcements, including a partnership with Open Innovation AI to deliver 30,000 GPUs of consumption to the Middle Eastern market. Integrating Nscale’s powerful GPU infrastructure with Open Innovation AI’s orchestration, data science tools and frameworks. Additionally, Nscale recently acquired Kontena, a leader in high-density modular data centres and AI data centre solutions, further enhancing its ability to provide high-performance, cost-effective AI infrastructure to meet the growing demands of the generative AI market.