NAVER and NVIDIA plan gigawatt-scale sovereign AI factory in South Korea using the NVIDIA DSX platform

On June 7, 2026, NAVER and NVIDIA announced a major sovereign AI infrastructure expansion: NAVER will build AI factories on the NVIDIA DSX platform at its GAK Sejong data center in South Korea, starting at 55 megawatts and scaling toward gigawatt capacity. The partnership makes NAVER one of Asia's most ambitious sovereign AI compute operators and the first Korean company to join NVIDIA's Nemotron Coalition.

What's new

Key specifics from the announcement:

• Scale: GAK Sejong starts at 55 MW, expands to 200 MW by 2028, then targets gigawatt-scale operations
• Technology: Built on NVIDIA DSX MaxLPS software and DSX OS — NVIDIA's full-stack platform for designing, building, and operating AI factories end-to-end
• Models: NAVER will fine-tune NVIDIA Nemotron 3 Ultra to advance its next-generation HyperCLOVA X models; will use NVIDIA Cosmos to develop a Seoul World Model
• Agentic AI: NAVER plans to launch an AI Agent Platform in Korea in H2 2026, powered by NVIDIA NemoClaw blueprints
• Nemotron Coalition: NAVER becomes the first Korean member of NVIDIA's Nemotron Coalition, contributing to open model pretraining, post-training, and reinforcement learning
• Export markets: The sovereign AI infrastructure will serve Korean enterprises and government, with Europe and the Middle East as future export markets

Context

The announcement was made by Haejin Lee, NAVER Founder and Chairman, and Jensen Huang, NVIDIA CEO, at an event aligned with NVIDIA's broader Asia-Pacific partnership push in the first week of June. On the same day, NVIDIA also announced similar AI factory partnerships with LG Group and Doosan Group in South Korea.

Jensen Huang framed the macro backdrop: "Useful AI has arrived, and demand for AI factories is extraordinary." Haejin Lee articulated NAVER's sovereign positioning: "NAVER is building sovereign AI infrastructure that can serve Korea's industries and global customers with trusted, high-performance AI."

NAVER's GAK Sejong facility is a next-generation hyperscale data center designed to power the company's growing AI and cloud services. The 2028 200 MW target and the gigawatt roadmap beyond that would place NAVER among the world's largest dedicated AI compute operators.

Why it matters

Sovereign AI infrastructure is an accelerating global pattern: nations and regional tech leaders are investing in national-scale compute to avoid dependence on US-hosted cloud AI. NAVER's commitment to gigawatt scale is the most ambitious Korean example to date. The export ambition — targeting Europe and the Middle East — signals that NAVER sees its infrastructure as a commercial product for other regions seeking trusted, non-US AI hosting. Joining the Nemotron Coalition also positions NAVER as an active contributor to the open model ecosystem, not just a consumer of NVIDIA compute. For enterprises in regulated industries across Asia and Europe, NAVER's sovereign AI factory could offer an alternative to AWS, Azure, or Google Cloud with stronger data-residency guarantees.