At the APEC Summit, among many country leaders, NVIDIA’s CEO Jensen Huang had the chance to be part of it and as such, he announced that Team Green will be supercharging South Korea’s AI infrastructure by working with partners and organizations to deploy more than 250,000 GPUs across sovereign clouds and AI factories.
The news came live as the Ministry of Science and ICT (MSIT) also announces its effort to deploy up to 50,000 of NVIDIA’s latest GPUs to bolster sovereign AI development, with initial rollouts including 13,000 of Blackwell + other GPUs through NVIDIA Cloud Partners such as NAVER Cloud, NHN Cloud, and Kakao Corp., with the deployment expected to expand through initiatives like the Korea National AI Computing Center.
Beyond national infrastructure, NVIDIA is also collaborating with leading Korean companies and institutions to advance AI-RAN and 6G technologies. Partners include Samsung, SK Telecom, ETRI, KT, LGU+, and Yonsei University, working together to create intelligent, low-power network solutions that offload GPU computing to base stations, reducing power usage and improving device performance.
Major Korean conglomerates are also constructing massive NVIDIA-powered AI factories to drive the next phase of industrial automation featuring Samsung building a semiconductor AI factory with more than 50,000 GPUs, using NVIDIA platforms like CUDA-X, Omniverse, and cuLitho to optimize chip production and develop next-gen robotics; SK Group designing a similar-scale AI factory focused on semiconductor R&D and digital twin technologies; SK Telecom to provide sovereign cloud infrastructure using NVIDIA RTX PRO 6000 Blackwell GPUs.
Meanwhile, Hyundai Motor Groupwill invest approximately $3 billion in AI infrastructure to enhance mobility, smart factories, and semiconductor development. Together with government agencies, they plan to establish the NVIDIA AI Technology Center, Hyundai Motor Group Physical AI Application Center, and multiple regional AI data centers. NAVER Cloud, meanwhile, will deploy over 60,000 GPUs for sovereign and physical AI projects, building inclusive AI models for industries like shipbuilding and security.
The Korean government’s Sovereign AI Foundation Models project, built with NVIDIA NeMo and Nemotron datasets, will enable research institutes, startups, and enterprises to develop large language models trained on local data. Participants include LG AI Research, NAVER Cloud, SK Telecom, NC AI, and Upstage, each contributing to a shared ecosystem for AI reasoning and speech-based agents.
In the realm of scientific research, NVIDIA is collaborating with KISTI to establish a Center of Excellence leveraging Korea’s sixth-generation HANGANG supercomputer. Using NVIDIA’s CUDA-Q platform and the new NVQLink open architecture, the partnership will explore hybrid quantum-GPU supercomputing, quantum error correction, and AI-driven scientific modeling through NVIDIA PhysicsNeMo.
To accelerate startup growth, NVIDIA and its partners are launching a new alliance under the NVIDIA Inception program. The initiative will provide Korean startups with access to NVIDIA cloud infrastructure, venture funding, and technical support, alongside collaboration with the Korea Ministry of SMEs and Startups through the N-Up AI incubation program.
