NVIDIA is extending its collaboration with NAVER through a large-scale AI infrastructure initiative centered on the NVIDIA DSX platform. The project begins with 55 megawatts of capacity and is designed to expand toward gigawatt-scale operations over time. Through the deployment, NVIDIA DSX will provide the foundation for full-stack AI environments capable of serving enterprises, industries, government organizations, and AI cloud customers. The companies said the platform is intended to streamline the creation, expansion, and operation of AI factories as demand for production-ready AI continues to grow.

The first phase of the rollout will support an expansion of NAVER’s GAK Sejong facility in South Korea. The hyperscale site will host a new cloud environment powered by NVIDIA DSX and optimized for accelerated computing workloads. According to the announcement, the facility incorporates automation systems, energy-efficiency measures, sustainability-focused capabilities, and disaster-response functions. The expansion is intended to strengthen support for NAVER’s growing AI and cloud services while providing infrastructure for large-scale digital operations.

The initiative arrives as AI applications increasingly transition from experimentation into operational deployment. Within that shift, AI factories are becoming an essential component for model training, post-training activities, and inference workloads. NVIDIA said its DSX platform will provide the technology stack required to support these functions while helping organizations scale AI production. The infrastructure is expected to provide a foundation for enterprises, manufacturers, public-sector organizations, and cloud customers seeking AI capabilities.

The infrastructure strategy also supports broader sovereign AI ambitions. NAVER said it is addressing demand for sovereign AI services through operations extending across Europe and the Middle East. By combining its hyperscale data center expertise with NVIDIA DSX technology, the company aims to deliver secure, high-performance digital services that align with regional regulatory and data-sovereignty requirements. The effort is also intended to support continued growth in NAVER’s AI data center business.

Beyond infrastructure, the partnership includes ongoing work in AI model development. NAVER is advancing its HyperCLOVA X family by refining the NVIDIA Nemotron 3 Ultra open model using proprietary data and training methods. The company has also joined the NVIDIA Nemotron Coalition and contributes to model development activities spanning pretraining, post-training, and reinforcement learning. Additional initiatives include plans for an AI Agent Platform in Korea and the development of a Seoul World Model built on NVIDIA Cosmos world foundation models.

📊 What This Means (Our Analysis)

This announcement illustrates how AI infrastructure is becoming a central requirement for deploying AI at production scale. NVIDIA’s focus extends beyond hardware alone, bringing together computing, software, operational tools, and infrastructure design into a single platform intended to support end-to-end AI deployment. The collaboration demonstrates how AI development increasingly depends on integrated ecosystems rather than isolated technologies.

The partnership also reflects the growing importance of regional AI capabilities built on locally controlled infrastructure. By combining AI factories, sovereign AI initiatives, cloud services, and model development within one framework, NVIDIA and NAVER are positioning their efforts around practical deployment and operational scale. The announcement highlights how infrastructure investment is emerging as a critical component of long-term AI growth strategies.

📌 Our Take: As AI adoption accelerates, the companies are placing large-scale infrastructure at the center of how future AI services will be built and delivered.