Nvidia deepens South Korea AI infrastructure push with SK Hynix, Naver and SK Telecom deals

Jensen Huang's Seoul visit produced new partnerships in memory, data centers and sovereign AI, broadening Nvidia's reach beyond earlier robotics and physical AI cooperation.

Summary

Nvidia expanded its South Korea push in AI infrastructure, memory and sovereign AI during CEO Jensen Huang's June visit to Seoul, adding new agreements with SK Hynix, Naver and SK Telecom while extending a broader set of industrial partnerships with groups including LG, Hyundai Motor Group and Doosan. SK Hynix signed a long-term technology pact with Nvidia to develop next-generation memory for four Nvidia computing platforms spanning AI infrastructure, personal AI and physical AI across servers, robotics, PCs and supercomputers. Huang said SK Hynix is Nvidia's largest memory supplier, that Nvidia already buys billions and billions of dollars of products from the company each year, and that those purchases are set to rise significantly. Naver is building an end-to-end Nvidia AI factory on the Nvidia DSX platform, and the expansion of its GAK Sejong plant would lift capacity from 55 MW to the gigawatt range. SK Telecom separately said it will build a gigawatt-class AI Cloud in South Korea based on Nvidia's DSX system, with the first AI data center expected to launch in 2027 and provide sovereign, physical and agentic AI capabilities. The announcements show Nvidia pushing beyond hyperscalers toward regional AI builders as countries and companies seek domestic computing capacity and tighter control over AI infrastructure. The visit also highlighted South Korea's role in robotics, gaming and physical AI. Huang said opportunities in the country are significant in robotics, physical AI and gaming, and his itinerary included meetings with senior business leaders as well as public appearances tied to esports, gaming and university events. Financial details of the new alliances were not disclosed.

Terms & Concepts
  • sovereign AI: AI infrastructure and services built to keep a country's data, computing capacity and model development under domestic control.
  • physical AI: AI applied to machines and robots operating in the real world.
  • AI factory: A large-scale computing facility designed to train, deploy and run AI workloads.