SK Telecom Debuts On-Demand GPUaaS AI Service

South Korea-based wireless service provider SK Telecom has unveiled its GPU-as-a-Service (GPUaaS) offering that the carrier is billing as an on-demand artificial intelligence (AI) cloud service that allows users to access graphics processing units (GPUs) with flexible pricing options.

“The GPUaaS allows businesses to choose the number and duration of GPUs depending on the scale or purpose of AI services, and can configure customized packages, such as standalone servers, firewalls and dedicated lines,” the SK Telecom said in a Jan. 13 announcement.

Pricing for the service is flexible depending on the length of the contract period, the number of GPUs being used, and the form of prepaid billing, the company said.

SK Telekom developed the new service at its AI Data Center in Gasan, Seoul through an investment in Lambda, a San Jose, Calif.-based provider of cloud services that specializes in GPUs and high performance computing services.

That investment agreement announced in August 2024 includes creation of a large-scale NVIDIA GPU cluster to support AI cloud business opportunities in South Korea.

“A strategic partnership with Lambda will strengthen our competitiveness in AI cloud services and develop greater business opportunities in Korea,” said Kim Kyeong-deog, vice president and head of Enterprise Business Division at SK Telecom when terms of the deal were set.

The latest service offering also builds on SK’s October 2024 launch of its SKT Enterprise Cloud Manager, which manages cloud services using GPUs.

“By utilizing the ‘AI Cloud Manager’ launched last October, GPU resources can be managed more efficiently,” SK Telecom explained earlier this month. “The AI Cloud Manager is a solution that manages numerous GPU resources as if they were a single computer to maximize GPU performance and support reducing the learning time required for AI development.”

“The introduction of GPUaaS marks the first instance of our AI Data Center business actively reaching out to customers,” commented Kim Myong-gook, who heads the SK Telecom GPUaaS business. “This is a crucial step in positioning SK Telecom as a major force in the AI infrastructure arena.”

Underlying the new GPUaaS service are NVIDIA H100 Tensor Core GPUs, with plans to introduce the NVIDIA H200 Tensor Core in the first quarter of this year. That coming move will “create an environment where domestic companies can develop AI technology and services more quickly and also to boost the expansion of GPUaaS customers,” SK Telecom said.