SK Telecom announced its latest strategy to build an 'AI infrastructure highway' and become an AI hub in the Asia-Pacific region

South Korean telecom operator SK Telecom recently announced a strategic plan to become an "AI hub" in the Asia-Pacific region by expanding into artificial intelligence data centers (AIDC), GPU-as-a-service (GPUaaS) and edge AI.

South Korean telecom operator SK Telecom (SK Telecom) recently announced a strategic plan to expand its approach to artificial intelligence data centers (AIDC), GPU-as-a-service (GPUaaS) and edge artificial intelligence (Edge). AI) field, and strive to become an "AI hub" in the Asia-Pacific region.

As the first step in this strategy, SK Telecom will build multiple "hyperscale AIDCs" with a power capacity of more than 100 MW in South Korea, and plans to expand to gigawatts or higher in the future. With this initiative, the company hopes to become an AIDC hub in the Asia-Pacific region.

SK Telecom announced that the data centers will be powered entirely by renewable energy sources such as hydrogen, solar, and wind, and will expand its services to the global market via submarine fiber optic cables.

The company plans to launch a pilot platform in Pangyo, South Korea, in December this year, which will integrate three types of liquid cooling solutions – direct liquid cooling, immersion cooling and precision liquid cooling – while also being equipped with advanced technologies such as AI semiconductors such as SK hynix HBM, GPU virtualization solutions, and AI energy optimization.

The company also plans to launch a cloud-based GPU-as-a-Service (GPUaaS), which aims to transform GPUaaS data centers into AIDC to efficiently address the current GPU supply shortage.

SK Telecom plans to launch NVIDIA H100 Tensor Cores in December this year in collaboration with its partner Lambda GPU service, and is expected to launch a service with NVIDIA H200 Tensor Core GPUs in March next year.

"In this way, we hope to enable enterprises to easily develop AI services at a lower cost without having to purchase more GPUs, ultimately supporting the thriving AI ecosystem in South Korea," the company said. ”

Finally, the company plans to launch "edge AI", which it claims can bridge the gap between AIDC and device-side AI, and is an infrastructure that combines mobile communication networks and AI computing. Edge AI offers advantages over large AIDC in terms of reduced latency, enhanced security, and improved privacy protection.

The company is working on various proof-of-concept projects in areas such as healthcare, AI robotics, and AI surveillance cameras.

Yoo Young-sang, CEO of SK Telecom, said, "Until now, competition in the telecom infrastructure space has revolved around connectivity such as speed and capacity, but now is the time to change the model of network development. In the coming 6G era, we will witness the deep integration of communications and artificial intelligence and will shape the next generation of AI infrastructure. ”

SK Telecom has clearly made artificial intelligence a priority. At this year's Mobile World Congress (MWC), the company (like most attendees) shined in the AI space, optimistic that telecom operators will be the main drivers of future growth.

During the MWC, Lee, Vice President and Head of Future R&D at SK Telecom In an interview with industry media, Jong-min said: "I think telcos will go beyond connectivity and embed intelligence into devices in the future. Therefore, we will provide AI-based services and technologies in the future. In the past, all devices were connected with the Internet of Things (IoT), but in the future, they will become 'connected to everything' (AI of Things), i.e. AI will be embedded in the device. ”