NAVER and NVIDIA Sign Major Infrastructure Partnership to Build Sovereign AI Factories and Gigawatt Scale Capacity for Enterprise and Government Customers
NAVER has signed a major infrastructure partnership with NVIDIA to build up sovereign AI in enterprise and government segments. Both have announced the collaboration to construct highly scalable AI factories based on the NVIDIA DSX platform, initially with a 55MW expansion at NAVER GAK Sejong data center, and with a roadmap for gigawatt scale of capacity for an AI Factory of the Future.
This industrial grade infrastructure has been developed to meet increased demand globally for locally hosted AI processing in order to minimize costs per token. The NVIDIA DSX platform contains both hardware and software to facilitate the swift building of the AI factories by NAVER to host these colossal GPU clusters. The company expects to achieve a secure high performance alternative for its enterprise and government customers whom need to maintain the stringent data sovereignty and local laws to address regulatory needs.
Useful AI has arrived, and demand for AI factories is extraordinary. NAVER is building AI factory infrastructure that will serve its companies, developers and industries. With NVIDIA DSX, we can help Korea scale sovereign intelligence infrastructure for the agentic era from AI agents to AI factories and physical AI. says Jensen Huang founder and CEO of NVIDIA.
This statement illustrates the significance of the project, that the infrastructure will provide capacity beyond processing raw power for agents and physical AI.
NAVERs latest generation of its GAK Sejong hyperscale data center will be the site of immediate development of the AI factories. The data center has been designed to be highly automated and capable of extremely high density accelerated computing. Energy efficiency, disaster recovery, and cooling are optimized. From this site, NAVER will begin exporting Sovereign AI cloud services internationally and across Europe and the Middle East, two key regions where strict laws exist to ensure localized control of data.
NAVER is building sovereign AI infrastructure that can serve Koreas industries and global customers with trusted, high performance AI. By building on the NVIDIA DSX platform, we can help customers move from AI experimentation to production scale AI factories that power models, agents and real world services. adds Haejin Lee founder and chairman of NAVER.
This adds weight to the fact that they will enable customers to transition from experimentation with AI to stable, industrial production through complex modeling at profitable margins.
The collaboration will span deeper than infrastructure and expand into software. NAVER will train its own Hyper CLOVA X models, using proprietary data in the refinement process, but also will use the NVIDIA Nemotron 3 Ultra open model to tailor systems for Korean business users and global enterprise clients. NAVER is the first Korean company to be named to the NVIDIA Nemotron Coalition which will help support open source training of models, and research in reinforcement learning.
The use of NAVERs street view data and spatial mapping data will construct the Seoul World Model, based on the NVIDIA Cosmos world foundation. An AI agent platform will be deployed in Korea later in the year using NVIDIA NemoClaw blueprints to provide AI assistants that are autonomous. Software in use will also include NVIDIA DSX MaxLPS which aims to optimize tokens per MW at low power and DSX OS to provide unattended multi tenancy and health of the entire infrastructure.
