Nvidia and SK Group have announced a groundbreaking AI project valued at over $500 billion, which includes the construction of a 2-gigawatt AI data center by SK Telecom utilizing Nvidia’s Vera Rubin chips and SK Hynix HBM4 memory. This initiative is part of a broader strategy, as SK Telecom plans to expand its AI data center capacity to up to 15 gigawatts by 2035, aligning with Korea’s efforts to advance domestic AI capabilities in areas such as robotics and sovereign computing. The partnership also focuses on joint research to co-develop next-generation AI factory architectures, enhancing infrastructure deployment and establishing SK Telecom as an NVIDIA Cloud Partner.
Nvidia: Nvidia is a technology company focused on graphics processing units and AI computing solutions. It recently announced a major collaboration with SK Group to develop AI infrastructure in South Korea, including the use of its Vera Rubin chips in new data centers.
SK Group: SK Group is a South Korean conglomerate with operations spanning semiconductors, telecommunications, and energy. It is partnering with Nvidia on joint research and deployment of next-generation AI factory architectures through its subsidiaries.
SK Hynix: SK Hynix is a South Korean semiconductor manufacturer specializing in memory products such as high-bandwidth memory. It is supplying HBM4 memory components for the AI data centers being developed in the Nvidia and SK Group project.
SK Telecom: SK Telecom is South Korea’s leading telecommunications provider and a key subsidiary of SK Group. It is spearheading the construction of gigawatt-scale AI data centers using Nvidia’s DSX platform, with the first facilities targeted for operation in 2027.
Expansion: Korean firms are advancing domestic AI capabilities to support robotics, industrial physical AI, and sovereign computing needs.
Partnership: Nvidia and SK Group are pursuing joint research to co-develop next-generation AI factory architectures beyond initial infrastructure deployment.
Infrastructure: SK Telecom is positioning itself as an NVIDIA Cloud Partner to deliver AI cloud services using full-stack reference architectures for improved performance and efficiency.