NAVER Builds AI Infrastructure With NVIDIA to Meet Surging Global Demand
NAVER will construct AI factories using NVIDIA's DSX platform at gigawatt scale, beginning with a 55-megawatt expansion at its GAK Sejong data center in South Korea. The partnership lets NAVER serve both domestic industries and global cloud customers with sovereign AI infrastructure.
The move reflects a shift in how companies approach AI. As generative AI and LLM applications move from experimentation to production, AI factories-facilities dedicated to training, post-training, and inference-are becoming operational necessities rather than research projects.
What NAVER Gets From DSX
NVIDIA's DSX platform is designed as an end-to-end stack for AI factories. It includes chips, systems, software, facilities design, and partner technologies. For product teams, this means a standardized blueprint rather than building infrastructure from scratch.
Two software components matter most. NVIDIA DSX MaxLPS achieves lower token costs by maximizing throughput per megawatt. NVIDIA DSX OS handles lifecycle management, runtime operations, health automation, and multi-tenant management across the facility.
NAVER's GAK Sejong facility was already built for high-density GPU computing with advanced automation and disaster-response systems. The DSX integration adds a unified operating layer.
The Model Development Angle
NAVER is fine-tuning NVIDIA's Nemotron 3 Ultra open model with proprietary Korean-language data and training methods. The result is HyperCLOVA X-a model designed for Korean and global enterprise customers.
NAVER also joined NVIDIA's Nemotron Coalition, contributing to open model development across pretraining, post-training, and reinforcement learning. This positions the company as both a builder and contributor to the broader AI ecosystem.
The company plans to launch an AI Agent Platform in Korea in the second half of 2026, powered by NVIDIA NemoClaw blueprints. It's also developing a Seoul World Model using street-view data and spatial modeling on NVIDIA's Cosmos foundation models.
Why Sovereign Infrastructure Matters
NAVER already operates infrastructure across Europe and the Middle East. The DSX expansion lets the company offer government and enterprise customers a trusted alternative for secure, high-performance services that meet local data-sovereignty and regulatory requirements.
For product teams, this means NAVER can serve customers who need AI capabilities but face restrictions on where data can be processed or stored. That's a competitive advantage in regulated industries.
What This Means for Product Development
The infrastructure investment directly supports NAVER's ability to offer AI agents and automation services at scale. Product teams building on NAVER's platform will have access to lower-cost inference, faster model serving, and standardized operational practices.
The gigawatt-scale plan suggests NAVER expects sustained demand for AI compute. That's a signal about where the market is heading-away from occasional model training runs and toward always-on AI services.
Your membership also unlocks:
