Overview
- Announced on the APEC sidelines, the plan was unveiled after Jensen Huang met President Lee Jae Myung and leaders of Samsung, SK Group, Hyundai and Naver.
- Roughly 50,000 GPUs each are designated for Samsung, SK and Hyundai, 60,000 for Naver Cloud and about 50,000 for government use in national AI infrastructure.
- Samsung will deploy 50,000 GPUs to build an AI Megafactory for automated chipmaking and in‑house models, using Nvidia’s Omniverse and a lithography platform adaptation Nvidia says boosts performance twentyfold.
- SK Group plans a manufacturing AI cloud and an AI factory powered by 50,000 GPUs, while SK Telecom signed an MoU with Nvidia, Samsung and ETRI to develop AI‑RAN technologies for 6G networks.
- South Korea and Nvidia will collaborate on sovereign foundation models and the National AI Computing Center, with financial terms and delivery schedules undisclosed and export constraints remaining a background consideration.