SK Telecom bets big on AI: Plans for AI 'superhighway' across Asia Pacific

SK Telecom bets big on AI: Plans for AI 'superhighway' across Asia Pacific

SK Telecom CEO Yoo Young-sang during his keynote at the SKT AI Summit

SK Telecom has unveiled plans to build an “AI infrastructure superhighway” which includes constructing gigawatt-scale data centres across Asia and the Pacific.

The plans were presented by CEO Ryu Young-sang during the company’s AI Summit event, with the Korean operator planning to develop AI data centres, a cloud-based GPU service, and edge AI solutions with partners in Korea and abroad.

Subscribe today for free


“So far, the competition in telecommunications infrastructure has been all about connectivity, namely speed and capacity, but now the paradigm of network evolution should be changed,” said Ryu. “The upcoming 6G will evolve into a next-generation AI infrastructure where communication and AI are integrated.”

Kicking off SK Telecom’s AI infrastructure superhighway plans will see the launch of an AI data centre testbed in Pangyo, Korea, in December.

The testbed will provide three types of liquid cooling and AI semiconductors to provide potential partners with opportunities to “observe and experience” the solutions a future AI data centre would employ.

Beyond the testbed, the operator plans to launch gigawatt-level AI data centres across the region. No details on locations of specific sites were disclosed, however, SK Telecom said they will use new renewable energy sources such as hydrogen, solar and wind power, and can easily support global markets through submarine cables.

The operator will also launch a GPU-as-a-Service (GPUaaS) by converting one of its existing data centres to provide cloud-based access to customers looking to run AI training and inference workloads.

The Gasan-based facility will be repurposed to revolve what the operator claims is a domestic shortage of GPUs.

Starting in December, SK Telecom will launch the service with Nvidia H100 GPUs through a partnership with US-based AI cloud provider Lambda.

The telco will then expand the service in March 2025, providing customers with access to Nvidia’s new H200 GPUs.

“Through the [GPUaaS], SKT aims to enable companies to develop AI services easily and at a lower cost, without needing to purchase their own GPUs, ultimately supporting the vitalisation of Korea’s AI ecosystem,” the company said.

The third rung of SK Telecom’s AI infrastructure superhighway sees the creation of edge AI services to combine mobile communication networks and AI computing.

The operator is conducting several edge-focused proof of concept projects, including healthcare, robotics, and CCTV to uncover potential specialised edge AI services.

SK Telecom said it aims to “lead the global standardisation of edge AI and collaborate on advanced technology research while working towards the transition to 6G AI infrastructure.”

RELATED STORIES

SK Telecom, Samsung use AI to optimise 5G base station performance

SK Telecom enhances AI capabilities with Lambda partnership

Gift this article