NVIDIA has officially brought another country under its closed ecosystem of AI infrastructure, this time joined by South Korea, as the firm announced at the APEC Summit.
The major partnership will see the deployment of more than 250K of NVIDIA GPUs across sovereign clouds and AI factories with the help of partners and industrial players. Specificailly, it’ll begin with the Ministry of Science and ICT (MSIT) confirmed plans to deploy up to 50,000 NVIDIA GPUs to advance sovereign AI development. The effort will start with an initial rollout of 13,000 NVIDIA Blackwell and other GPUs through partners including NAVER Cloud, NHN Cloud, and Kakao Corp., contributing to national initiatives such as the establishment of Korea’s National AI Computing Center.
NVIDIA is also collaborating with Korean telecom operators and research institutions on AI-RAN and 6G technologies. Partners such as Samsung, SK Telecom, KT, LGU+, ETRI, and Yonsei University are developing low-power, intelligent networks that can shift GPU computation to base stations, improving performance and extending device battery life.
Korea’s industrial leaders are investing heavily in NVIDIA-powered AI factories to drive advanced manufacturing. Samsung is building a semiconductor AI factory with over 50,000 GPUs, using NVIDIA Nemotron datasets, CUDA-X, cuLitho, and Omniverse to create digital twins and improve production efficiency. SK Group is constructing an AI factory of similar scale to enhance semiconductor research and AI agent development, while SK Telecom is introducing sovereign cloud infrastructure featuring NVIDIA RTX PRO 6000 Blackwell GPUs.
Meanwhile, Hyundai Motor Group will be investing around $3 billion to develop AI capabilities across smart factories, mobility, and on-device semiconductors. The two will jointly establish the NVIDIA AI Technology Center and Hyundai Motor Group Physical AI Application Center as part of Korea’s national AI cluster initiative. Meanwhile, NAVER Cloud will expand its NVIDIA infrastructure to over 60,000 GPUs, focusing on sovereign AI models for industries such as shipbuilding and security.
To support Korea’s Sovereign AI Foundation Models project, NVIDIA is partnering with LG AI Research, SK Telecom, NAVER Cloud, NC AI, and Upstage to develop locally trained large language models using NVIDIA NeMo and Nemotron datasets. The goal is to strengthen national AI capabilities while fostering collaboration between enterprises, startups, and research institutions.
In scientific research, the Korea Institute of Science and Technology Information (KISTI) is working with NVIDIA to establish a Center of Excellence powered by Korea’s sixth-generation supercomputer, HANGANG. Through NVIDIA CUDA-Q and NVQLink, KISTI will advance research in quantum error correction, hybrid quantum-GPU systems, and AI-based scientific computing using the open-source PhysicsNeMo framework.
NVIDIA is also launching a new startup alliance under its Inception program to accelerate AI innovation in Korea. Supported by SK Telecom, IMM Investment, Korea Investment Partners, and SBVA, the alliance will provide startups with access to NVIDIA’s infrastructure, venture funding, and technical resources. NVIDIA will also support the Ministry of SMEs and Startups’ N-Up AI incubation program to nurture the next generation of Korean AI companies.
