NVIDIA CEO Jensen Huang has visited Seoul to meet with leaders from SK Group, LG Group, and Naver, announcing plans to expand cooperation in AI semiconductors, physical AI, robots, and data centers. Huang hinted at significant business expansion for Korea in these advanced technology sectors during his visit. This marks his first return to Korea in seven months.
NVIDIA is vertically integrating its supply chain by moving from a pure chip supplier to a foundational partner in Korea’s industrial infrastructure. By locking in gigawatt-scale data center capacity and HBM4 memory supply, NVIDIA secures the hardware and power required to sustain the Blackwell and Rubin GPU cycles through the late 2020s. For the streaming and media ecosystem, this accelerated infrastructure buildout enables the localized, sovereign AI models necessary for high-scale video inference and personalized content generation. Watch for the 2027 activation of SK Telecom’s first 1GW AI factory as a benchmark for large-scale GPU deployment efficiency.
The visit by Jensen Huang in June 2026 marks a decisive pivot toward 'Physical AI,' where NVIDIA uses its Cosmos world foundation model to simulate and train robotics in digital twin environments. Per NVIDIA’s official June 2026 briefings, the company is partnering with LG Electronics to integrate high-performance sensing modules from LG Innotek into the Isaac robotics platform. This collaboration aims to commoditize humanoid and logistics robots, using LG's vast manufacturing data to refine NVIDIA’s GR00T AI models. These robots are intended for deployment across global automotive and consumer electronics factory floors, positioning South Korea as a primary laboratory for autonomous manufacturing.
On the infrastructure front, the 'AI Factory' concept is driving massive capital expenditures in the region. According to reporting from The Elec and NVIDIA in early June 2026, Naver has become the first South Korean company to join the NVIDIA Nemotron Coalition, a global consortium focused on open-source large language model development. By fine-tuning the 550-billion-parameter Nemotron-3 Ultra with proprietary Korean data, Naver is positioning its HyperCLOVA X model as a leading sovereign AI alternative for government and enterprise clients in Asia and the Middle East. This alignment helps mitigate over-reliance on US-centric models while utilizing NVIDIA's full-stack DSX platform for deployment.
Furthermore, the multi-year memory supply agreement with SK Hynix is a response to the persistent global shortage of High Bandwidth Memory (HBM). Per analyst consensus from NH Investment & Securities in June 2026, HBM demand is currently growing at nearly double the rate of supply. By securing a preferential supply of 12-layer HBM4 products, NVIDIA protects its Vera Rubin architecture rollout scheduled for later in 2026. This deal effectively sidelines competitors from immediate premium memory allocations, ensuring that NVIDIA’s largest cloud customers—including AWS and Microsoft Azure—can maintain their aggressive AI infrastructure roadmaps.
Add StreamingMeme as a preferred source on Google to see more of our streaming news at the top of your Search results.
Add as preferred source