On June 8, 2026, South Korean internet leader NAVER and NVIDIA jointly announced a full-stack AI strategic partnership. Under the collaboration, NAVER will expand its sovereign AI infrastructure based on NVIDIA’s DSX platform and build a large-scale AI computing cluster for the Korean and global markets.
Under the agreement, NAVER will begin deployment at its GAK data center in Sejong with an initial 55-megawatt AI computing capacity, followed by phased expansion toward gigawatt-scale infrastructure. The project is expected to become South Korea’s first gigawatt-scale AI factory. The infrastructure will serve Korean enterprises, manufacturers, government agencies, and global AI cloud customers by providing cost-effective AI training and inference services. As a standardized AI factory solution, NVIDIA’s DSX platform will provide end-to-end technical support across design, deployment, and operations.
In AI model development, NAVER will use NVIDIA’s full-stack AI platform to accelerate the development of the next-generation HyperCLOVA X Korean large language model. NAVER will also fine-tune NVIDIA’s open-source Nemotron 3 Ultra model by combining Korean-language data with industry expertise to build localized specialized models for the Korean and global enterprise markets. In addition, NAVER has officially joined the NVIDIA Nemotron Alliance, becoming the first Korean company to do so. The company will participate in the development of pre-training, post-training, and reinforcement learning technologies for open-source large language models.
The partnership also extends to physical AI. NAVER will use NVIDIA’s Cosmos platform to optimize its “Seoul World Model,” enabling high-precision digital replication and simulation of urban spaces. NAVER plans to launch an AI agent platform in South Korea in the second half of 2026, offering enterprises AI service capabilities for autonomous decision-making and automated execution.