NVIDIA and South Korea’s Ministry of Science and ICT Expand Cooperation on Sovereign AI

NVIDIA said Nov. 9, 2026, that it would expand cooperation with South Korea’s Ministry of Science and ICT (MSIT) on sovereign AI. The areas of cooperation include AI semiconductors, data center infrastructure, talent, startups, models, agents and security.

NVIDIA said the cooperation includes laying the groundwork for South Korean NPU and XPU companies to consider joining the NVLink Fusion ecosystem. How South Korean chip development connects with NVIDIA’s AI infrastructure will be a technical and business focus if participation takes shape.

The announcement spans not only AI semiconductors and data centers but also talent, startups, models, agents and security. The focus, therefore, is not on an individual GPU procurement or a single robotics deployment, but on including multiple elements that support AI development and operations in South Korea in the scope of cooperation. The announcement calls for expanded cooperation, but the information provided does not specify investment amounts or the scale of implementation in each area. The clearest feature of the arrangement is the connection between South Korean NPU and XPU companies and NVLink Fusion. NVIDIA said it would lay the groundwork for these companies to “consider” joining the ecosystem; it did not announce a decision to join or an implementation of a connection. The announcement presents the relationship between South Korean chip development and NVIDIA’s infrastructure as a framework for exploring the possibility of connecting them, rather than as an exclusive choice. However, the excerpt provided does not specify the scope of compatibility or the conditions for connection, so technical integration cannot be treated as an established fact. In an article published the same day, Physical AI News covered NVIDIA’s explanation that it would support Physical AI applications in semiconductor manufacturing, robotics and autonomous vehicles in South Korea’s manufacturing and engineering sectors “NVIDIA Promotes Physical AI in South Korea’s Semiconductor, Robotics and Autonomous Driving Industries”. Unlike that account of industrial applications, this announcement describes ecosystem-level cooperation that also includes talent, models and security, in addition to semiconductors and computing infrastructure. Taken together, the two announcements do not establish the scale of specific facilities or implementation plans. The announcement also differs in focus from a same-day article on the outlook for South Korea’s GPU supply “South Korea Sees No Disruption to Contracted NVIDIA GPU Supply, Vice Science Minister Says”. The status of supply contracts and the scope of sovereign AI cooperation should be considered separately. The industrial significance lies in addressing, within the same cooperation effort, both South Korean companies’ development of domestic NPUs and XPUs and the possibility of joining NVIDIA’s ecosystem. If participation advances, the issues will extend beyond chip selection to include the conditions for connecting with data center infrastructure and models. The next points to watch are which companies are considering participation, the technical scope of connection with NVLink Fusion, the scale and timing of data center investment, and the division of roles in models and security. These implementation conditions cannot be confirmed from the excerpt of the announcement provided.

NVIDIA said the cooperation covers South Korea’s sovereign AI ecosystem as a whole. It also includes laying the groundwork for South Korean NPU and XPU companies to consider joining the NVIDIA NVLink Fusion ecosystem. The excerpt of the announcement provided does not specify the process for considering participation, the companies involved, investment amounts, or the scale and operating schedule of data centers.