SK Telecom announced on August 1 that it has launched Korea’s largest and most powerful GPU cloud service, named GPU-as-a-Service (GPUaaS) based on Nvidia’s latest AI semiconductor, the Blackwell B200 GPU.
The new service comprises more than 1,000 Nvidia Blackwell B200 GPUs assembled into a single cluster, representing a significant upgrade from previous H100-based services. This infrastructure is optimized for large-scale AI model training and inference, providing high-performance GPU resources that address the growing demand for AI infrastructure in Korea. It is expected to enhance the competitiveness of the domestic AI industry and support the development of advanced AI capabilities.
SK Telecom has installed this service within its AI data center in Gasan and named the cluster ‘Haein (海印).’ The name ‘Haein’ symbolizes wisdom in the digital age and is inspired by Haeinsa Temple, where the Tripitaka Koreana is stored. The Haein cluster was selected as part of the Ministry of Science and ICT’s ‘AI-Computing Resources Utilization Enhancement Project,’ serving as a core infrastructure for developing foundation AI models in Korea.
This cluster is a product of SK Telecom’s strategic AI initiatives and global collaborations. The company signed a $200 million investment agreement with Penguin Solutions, a U.S.-based high-performance computing (HPC) specialist, and has been fostering strategic cooperation since. Simultaneously, SK Telecom partnered with Supermicro, a leading AI server manufacturer, to quickly secure GPU server supplies, ensuring service stability and speed.
Technologically, SK Telecom applied its proprietary solutions, utilizing the ‘Petasus AI Cloud,’ a virtualization platform that flexibly divides and manages GPU resources. Additionally, the AI Cloud Manager, an AIOps platform, supports efficient AI development and operation, enhancing productivity and ease of use.
Following the opening of Lambda’s Korea region last December, the launch of the B200-based GPUaaS marks SK Telecom’s strengthened position as a global AI infrastructure provider. The company plans to continue expanding its GPUaaS business to support Korea’s AI industry and enhance its competitiveness in the global market.
Kim Myung-guk, head of SK Telecom’s GPUaaS division, stated, “The Haein cluster, built with the latest B200 GPUs, is Korea’s largest AI infrastructure and a pivotal turning point for our customers and the nation’s AI capabilities. SK Telecom will lead the way in building AI infrastructure and opening Korea’s AI high-speed highway.”
