Rotating Data
Explore the map
Home/Deal Room/NAVER and NVIDIA announce gigawatt-scale Korean AI factories

Data centers · Evidence brief

NAVER and NVIDIA announce gigawatt-scale Korean AI factories

NAVER + NVIDIA

Primary sourcePress release only2026-06-07
Deal structureNVIDIA DSX sovereign AI infrastructure expansion / AI cloud buildout
GeographySejong, South Korea, with planned global expansion across NAVER Cloud regions
Infrastructure layersData centers + cloud capacity + chips / compute + networking + power infrastructure
Disclosed scaleGigawatt-scale NVIDIA DSX AI factory plan, starting with 55 MW of AI infrastructure expansion at NAVER GAK Sejong and planned ramp to 100 MW overseas in 2027 and 200 MW by 2028

Why it matters

The strategic read

Adds another primary-source Korean DSX buildout beside SK Telecom. NAVER already operates hyperscale cloud and GPU clusters, so the announcement turns sovereign AI, enterprise AI cloud, agentic services, physical AI and regional data-sovereignty demand into a concrete AI-factory capacity ramp starting at GAK Sejong and pointing toward gigawatt-scale NVIDIA-aligned infrastructure.

Source-supported terms

What was disclosed

  1. 01

    NVIDIA and NAVER announced on June 7, 2026 that NAVER will build NVIDIA DSX-based AI factories at gigawatt scale.

  2. 02

    NVIDIA said NAVER will expand sovereign AI infrastructure starting at 55 MW with plans to move to gigawatt scale using the NVIDIA DSX platform.

  3. 03

    The first DSX-powered cloud expansion will add high-density NVIDIA accelerated computing at NAVER GAK Sejong, a hyperscale data center in Sejong, South Korea.

  4. 04

    DCD reported the 55 MW capacity is due online in 2027, with NAVER then aiming to expand overseas AI infrastructure to 100 MW within 2027 and 200 MW by 2028.

  5. 05

    NAVER plans to use the infrastructure for Korean sovereign AI, global AI cloud customers, next-generation HyperCLOVA X models, Seoul World Model development and agentic AI services.

  6. 06

    NVIDIA said NAVER is the first Korean company to participate in the NVIDIA Nemotron Coalition and plans to launch an AI Agent Platform in Korea in the second half of 2026.

  7. 07

    NVIDIA DSX is described as a codesigned end-to-end AI factory stack spanning chips, systems, software, facilities and partner technologies to lower token cost and speed time to production.

  8. 08

    DCD noted NAVER completed a 4,000 NVIDIA B200 GPU cluster in South Korea in January 2026 and that GAK Sejong is planned to reach 270 MW of capacity and up to 600,000 servers.

  9. 09

    Public materials do not disclose capex, GPU count for the new 55 MW phase, utility service agreements, PPAs, financing, EPC awards, data-center construction packages, exact overseas sites, anchor customer contracts or final delivery milestones beyond the early capacity targets.

Disclosure boundary

What is known—and what is not

Primary NVIDIA Newsroom announcement plus DCD near-primary capacity detail. Public details disclose the parties, DSX platform, gigawatt-scale intent, initial 55 MW GAK Sejong expansion, early 2027/2028 ramp targets, sovereign and enterprise AI-cloud use cases, and NAVER model / agent platform links. Treat as a strategic AI-factory infrastructure announcement rather than a fully financed construction contract because detailed capex, power, financing, equipment, EPC and customer/offtake documents are not public.