NAVER and NVIDIA announced a major expansion of South Korea's sovereign AI infrastructure, starting at 55 megawatts and scaling toward gigawatt capacity. The partnership leverages NVIDIA's DSX platform to build end-to-end AI systems, with NAVER planning to launch an AI Agent Platform in the second half of the year powered by NVIDIA NemoClaw blueprints.

What Is NemoClaw and Why Does It Matter for Enterprise AI?

NemoClaw is NVIDIA's blueprint framework for building agentic AI systems, which are AI agents designed to autonomously complete complex tasks across enterprises. Unlike traditional chatbots that respond to individual queries, agentic systems can plan, execute, and adapt across multiple steps. NAVER's decision to build its AI Agent Platform on NemoClaw blueprints signals a shift toward production-ready agent deployment in Korea, where regulatory and data sovereignty requirements demand locally controlled infrastructure.

The expansion centers on NAVER's GAK Sejong data center, a hyperscale facility in Sejong, South Korea, designed for high-density computing with advanced automation and disaster-response capabilities. This infrastructure will power not only NAVER's own AI services but also serve enterprise and government customers seeking alternatives to foreign cloud providers.

How Is NAVER Building Its Sovereign AI Advantage?

• Model Development: NAVER is fine-tuning NVIDIA's Nemotron 3 Ultra open model with proprietary Korean data and training expertise to create HyperCLOVA X, a culturally fluent model for Korean and global enterprise customers.
• Agent Platform Launch: NAVER plans to deploy an AI Agent Platform in Korea during the second half of the year, powered by NemoClaw blueprints for autonomous task execution.
• World Model Innovation: NAVER is developing a Seoul World Model using proprietary urban street-view data and spatial modeling technology, built on NVIDIA Cosmos world foundation models for spatial reasoning.
• Coalition Participation: NAVER became the first Korean company to join the NVIDIA Nemotron Coalition, contributing to open model development across pretraining, post-training, and reinforcement learning.

The NVIDIA DSX platform itself is purpose-built as an end-to-end stack for AI factories, spanning chips, systems, software, facilities, and partner technologies. DSX MaxLPS software achieves the lowest token cost by maximizing token throughput per megawatt, while DSX OS provides lifecycle management, runtime operations, health automation, and multi-tenant AI factory management.

Why Is Sovereign AI Infrastructure Critical Right Now?

NAVER is serving sovereign AI demand with infrastructure across Europe and the Middle East, positioning itself as a trusted alternative for secure, high-performance digital services compliant with local regulatory and data-sovereignty requirements. This matters because enterprises and governments increasingly face pressure to keep sensitive data within national borders, making locally controlled AI infrastructure a competitive advantage.

The partnership builds on NVIDIA and NAVER's existing collaboration in advancing sovereign and physical AI models in Korea. By combining NAVER's hyperscale data center expertise with NVIDIA's full-stack AI platforms, the two companies aim to accelerate both model development and the broader AI data center business. The 55-megawatt starting point represents a significant commitment, with plans to scale toward gigawatt-level capacity as demand grows.

NAVER's AI Agent Platform launch in the second half of 2026 will be a critical test of whether NemoClaw blueprints can deliver production-ready agentic systems for Korean enterprises. The combination of sovereign infrastructure, locally optimized models, and agent-based automation positions NAVER to capture demand from organizations that need both AI capability and regulatory compliance in a single solution.