NAVER, Brookfield and NVIDIA on July 24, 2026 announced a planned expansion of NAVER’s NVIDIA DSX AI factory at its GAK Sejong hyperscale data center in Sejong, South Korea, lifting capacity to 200 megawatts.
The planned expansion would more than triple the initial buildout announced in June, which stood at 55 megawatts. NVIDIA said the expanded factory will run on roughly 100,000 GPUs built on the NVIDIA Vera Rubin platform.
The 200-megawatt site is also expected to feature NVIDIA Vera Rubin and Blackwell platforms under the NVIDIA DSX stack.
The announcement came during Korea President Jae Myung Lee’s AI Summit visit to San Francisco. NAVER and NVIDIA plan to complete the step from 55 megawatts to 200 megawatts by 2028.
New investments support that scale-up of the national AI factory infrastructure.
The factory will create a dedicated resource pool for emerging AI companies. It is set to supply the compute, software and support those firms need to develop and deploy competitive AI models and applications at scale.
NAVER, acting as an NVIDIA Cloud Partner, brings expertise operating hyperscale infrastructure on the full-stack NVIDIA AI platform.
This expansion builds on NAVER’s June commitment to extend GAK Sejong with NVIDIA DSX. That earlier plan already sketched a long-term path to gigawatt-scale sovereign AI infrastructure serving Korea’s enterprises, industries, government organizations and global AI cloud customers.
NAVER intends to push its NVIDIA AI infrastructure deployment toward 1 gigawatt and continues that path with the DSX platform.
The larger site also extends NAVER and NVIDIA’s work on open model development for agentic and physical AI. NAVER plans to launch an AI agent platform in Korea in the second half of the year.
That platform will run on NVIDIA Agent Toolkit software, including NVIDIA NemoClaw blueprints.
NVIDIA DSX MaxLPS software is described as maximizing token throughput per megawatt across the expanded plant. NVIDIA DSX OS is listed for lifecycle management, health automation, resiliency and multi-tenant AI factory management.
Together the partners frame the buildout as fuel for AI innovators inside Korea’s broader infrastructure push.
Readers tracking sovereign compute will see a concrete capacity jump at a single named Sejong campus, tied to Vera Rubin and Blackwell hardware and a second-half agent platform release. Full details appear in the NVIDIA account of the Korea partner announcements.
