NAVER and NVIDIA expand Korea AI factory to 200MW
A $10 billion funding package will scale the Korea AI factory and provide production-level computing to AI developers.
NAVER, NVIDIA and Brookfield have proposed a $10 billion expansion of an AI computing facility in South Korea, increasing its planned capacity from 55 megawatts to 200 megawatts by 2028.
The infrastructure would be built at NAVER’s GAK Sejong hyperscale data centre using NVIDIA’s DSX platform. It is intended to give AI developers in South Korea and the US access to production-scale computing for developing models, autonomous agents and AI-powered services.
Project financing has not yet been fully secured. Brookfield has signed a non-binding term sheet to provide up to $9 billion, while NVIDIA plans to invest $1 billion in NAVER. NVIDIA’s investment is subject to customary closing conditions and NAVER securing at least $9 billion in committed project financing. NAVER would cover any remaining costs.
The expanded facility is expected to use NVIDIA’s Blackwell and Vera Rubin computing platforms. NAVER also plans to establish a dedicated resource pool for emerging AI companies, providing access to computing infrastructure, software and technical support.
The project builds on NAVER’s longer-term plan to develop gigawatt-scale AI infrastructure serving South Korean companies, industries, government organisations and international cloud customers.
NAVER and NVIDIA also announced further cooperation on AI models and applications. NAVER said future HyperCLOVA X development would use NVIDIA’s NeMo 3 Ultra open models alongside its proprietary data and training methods.
The company plans to launch an AI agent platform in South Korea during the second half of 2026 using NVIDIA software. It is also developing a Seoul World Model from proprietary street-view and spatial data using NVIDIA’s Cosmos foundation models.
Why does it matter?
Large AI facilities are becoming part of national industrial policy because access to advanced computing can determine which companies can develop and deploy competitive models. A 200-megawatt facility could reduce South Korean developers’ reliance on overseas data centres and widen domestic access to scarce computing resources. However, it would not remove dependence on foreign technology suppliers, since the project is built around NVIDIA hardware and software, while most of the financing has not yet been finalised.
Would you like to learn more about AI, tech, and digital diplomacy? If so, ask our Diplo chatbot!