Naver partners with Nvidia to build gigawatt-scale AI factory
Translated from Korean, summarized and contextualized by DistantNews.
At a glance
- Naver and Nvidia will collaborate to build a gigawatt-scale AI infrastructure, four times the size of South Korea's largest data center.
- The project will utilize Nvidia's DSX platform to create an
Naver is partnering with Nvidia to construct a massive artificial intelligence (AI) infrastructure, aiming for gigawatt-scale capacity that dwarfs South Korea's current largest data center.
This collaboration goes beyond mere technology sharing. Naver and Nvidia plan to jointly develop business opportunities, including identifying global demand in Asia, the Middle East, and Europe, and exploring capital partnerships.
The initiative will leverage Nvidia's DSX platform, designed for building and operating "AI Factories." Naver plans to begin with a 55-megawatt (MW) infrastructure at its "Gak Sejong" hyperscale data center in the first half of next year. By 2028, the goal is to expand this capacity to 200 MW globally, a crucial first step towards a gigawatt-level AI factory.
AI Factories, as envisioned by Nvidia, are next-generation data centers that utilize large clusters of Graphics Processing Units (GPUs) to train AI models, infer data, and produce AI tokens and services. Unlike traditional data centers focused on storage and processing, AI Factories are designed as production hubs for AI services.
This partnership also aims to enhance data center operational efficiency and business viability by combining Naver's expertise in managing large GPU clusters and hyperscale data centers with Nvidia's DSX platform. Further collaboration includes integrating Naver's "Seoul World Model," built with spatial modeling and street view data, into Nvidia's "Omniverse" platform for spatial intelligence.
Naver's involvement in this project aligns with its expanding participation in the global AI ecosystem. The company recently joined the "Nvidia NeMo Guardrails Alliance," a consortium of leading AI firms. Naver intends to leverage its data and learning expertise to bolster the performance and global competitiveness of its own large language model, HyperCLOVA X.
Through this alliance, we can present a concrete alternative that allows each region and country around the world to build its own sovereign AI capabilities, which is very encouraging.
Originally published by Hankyoreh in Korean. Translated, summarized, and contextualized by our editorial team with added local perspective. Read our editorial standards.