Singtel, NVIDIA to Bring Sovereign AI to Southeast Asia

Singapore’s leading service provider chooses the NVIDIA AI platform to deliver generative AI to its regional customers.
by Joao Gomes

Asia’s lion city is roaring ahead in AI.

Singtel, a leading communications services provider based in Singapore, will bring the NVIDIA AI platform to businesses in the island nation and beyond.

The mobile and broadband company is building energy-efficient data centers across Southeast Asia accelerated with NVIDIA Hopper architecture GPUs and using NVIDIA AI reference architectures proven to deliver optimal performance.

The data centers will serve as sovereign national resources — AI factories that process the private datasets of companies, startups, universities and governments safely on shore to produce valuable insights.

Singtel’s first AI services will spin up in Singapore, with future data centers under construction in Indonesia and Thailand. From its hub in Singapore, the company has operations that stretch from Australia to India.

Trusted Engines of AI

The new data centers will act as trusted engines of generative AI. The most transformative technology of our time, generative AI and its ability to amplify human intelligence and productivity are attracting users worldwide.

Nations are creating large language models tuned to their local dialects, cultures and practices. Singtel sits at the center of such opportunities among Southeast Asia’s vibrant Chinese, Indian, Malay and other communities.

Singtel’s initiative supports Singapore’s national AI strategy to empower its citizens with the latest technology. The plan calls for significantly expanding the country’s compute infrastructure as well as its talent pool of machine learning specialists.

For businesses in the region, having a known, local provider of these computationally intensive services provides a safe, easy on-ramp to generative AI. They can enhance and personalize their products and services while protecting sensitive corporate data.

Taking the Green Path

Singtel is committed to democratizing AI and decarbonizing its operations.

Its latest data centers are being built with an eye to sustainability, including in the selection of materials and use of liquid cooling. They adopt best practices to deliver less than 1.3 in PUE, the power usage effectiveness metric for data center efficiency.

Singtel will use its Paragon software platform to orchestrate how the new AI applications work in concert with its mobile and broadband services. The combination will enable edge computing services like powering robots and other autonomous systems from AI models running in the cloud.

A Full-Stack Foundation

The company will offer its customers NVIDIA AI Enterprise, a software platform for building and deploying AI applications, including generative AI. Singtel will also be an NVIDIA Cloud Partner, delivering optimized AI services on the NVIDIA platform.

Because Singtel’s data centers use NVIDIA’s proven reference architectures for AI computing, users can employ its services, knowing they’re optimized for leading AI performance.

Singtel already has hands-on experience delivering edge services with NVIDIA AI.

Last May, it demonstrated a digital avatar created with the NVIDIA Omniverse and NVIDIA NeMo platforms that users could interact with over its 5G network. And in 2021, Singtel delivered GPU services as part of a testbed for local government agencies.

New AI Role for Telcos

Singapore’s service provider joins pioneers in France, India, Italy and Switzerland deploying AI factories that deliver generative AI services with data sovereignty.

To learn more about how Singtel and other telcos are embracing generative AI, register for a session on the topic at NVIDIA GTC. The global AI conference runs March 18-21, starting with a keynote by NVIDIA founder and CEO Jensen Huang.