en.Wedoany.com Reported - NAVER will expand its sovereign AI infrastructure, starting at 55 megawatts, with plans to scale to gigawatt levels using the NVIDIA DSX platform, enabling rapid design, construction, and expansion of full-stack end-to-end AI platforms serving enterprises, industries, and governments.
As artificial intelligence enters the production phase, AI factories have become critical infrastructure for training, post-training, and inference. Built on NVIDIA accelerated computing and the NVIDIA DSX platform, NAVER's AI factory will provide a sovereign foundation for South Korea, generating intelligence for enterprises, manufacturers, government organizations, and AI cloud customers.
The new DSX-based cloud will expand NAVER's GAK Sejong Data Center, a next-generation hyperscale facility located in Sejong City, South Korea, designed for high-density NVIDIA accelerated computing. GAK Sejong delivers resilient, energy-efficient operations through advanced automation, sustainability features, and robust disaster response capabilities, enabling NAVER to support large-scale digital services and innovation in South Korea and globally. NAVER also leverages its infrastructure in Europe and the Middle East to serve sovereign AI needs, aiming to combine hyperscale data center capabilities with NVIDIA DSX to provide governments and customers with secure, high-performance digital services that comply with local regulations and data sovereignty requirements.
NAVER will use the NVIDIA full-stack AI platform to power sovereign AI models and agentic AI services, while accelerating its AI data center business. The new AI cloud infrastructure builds on the partnership between NVIDIA and NAVER to advance sovereign AI and physical AI models in South Korea. By fine-tuning the NVIDIA Nemotron 3 Ultra open model using its proprietary data and training expertise, NAVER advances its HyperCLOVA X model, delivering more powerful and culturally fluent models for enterprise customers in South Korea and globally. The NVIDIA Nemotron-based HyperCLOVA X model also provides a platform supporting sovereign AI initiatives in Europe and the Middle East. NAVER is the first South Korean company to join the NVIDIA Nemotron Coalition, dedicated to advancing open model development in pre-training, post-training, and reinforcement learning, and plans to launch an AI agent platform powered by the NVIDIA NemoClaw blueprint in South Korea in the second half of this year. NAVER is also developing the Seoul World Model based on the NVIDIA Cosmos world foundation model, utilizing its proprietary city street view data and spatial modeling technology.
The NVIDIA DSX platform is an end-to-end co-designed stack for AI factories, encompassing chips, systems, software, facilities, and partner technologies, designed to minimize token costs and reduce time to first token. This platform will be combined with NAVER's expertise in building and operating large-scale proprietary GPU clusters and hyperscale data centers. NVIDIA DSX MaxLPS software achieves the lowest token cost by maximizing token throughput per megawatt, while NVIDIA DSX OS provides an operating system layer that supports vendor lifecycle management, consistent runtime operations, health automation, resilience, and multi-tenant AI factory management through open-source modular software.
This article is compiled by Wedoany. All AI citations must indicate the source as "Wedoany". If there is any infringement or other issues, please notify us promptly, and we will modify or delete it accordingly. Email: news@wedoany.com
