Naver announced on June 8 that it has reached an agreement with NVIDIA to jointly establish a gigawatt-scale global artificial intelligence (AI) factory. The collaboration aims to leverage NVIDIA's global infrastructure standards alongside Naver's cloud operations to position the company as a key player in the business-to-business (B2B) AI infrastructure market.
Naver described the partnership as an integrated collaboration that spans the entire value chain, from identifying global demand to capital cooperation.
Notably, Naver will participate as a core global partner, sharing both the successes and risks of the venture. The companies plan to initiate the AI factory project next year, starting with a capacity of 55 megawatts (MW) to capture the surging global demand for AI.
One gigawatt (GW) is approximately four times the maximum capacity of Naver's hyperscale data center, Gak Sejong, which can accommodate hundreds of thousands of NVIDIA's latest graphics processing units (GPUs) simultaneously.
Naver Chairman Lee Hae-jin and NVIDIA CEO Jensen Huang met at Naver's headquarters to discuss the specific roadmap for their ongoing projects and detailed cooperation strategies for entering global markets. Naver stated that both companies have already agreed on a direction to secure leadership in the AI infrastructure ecosystem beyond the Asia-Pacific region, extending to Europe and the Middle East.
Naver will utilize its hyperscale data center, Gak Sejong, as a base for operations. The company plans to begin operations with 55MW in the first half of next year, expanding to 100MW within the same year and reaching 200MW by 2028 to meet global demand. This is part of a phased roadmap aimed at establishing a gigawatt-scale infrastructure that will dominate the market.
Gak Sejong will serve as a base for Naver's AI 'HyperCLOVA X,' with its power infrastructure designed to support the operation of large GPU clusters from the outset. It is six times larger than Naver's first data center, Gak Chuncheon, with a maximum server capacity of 600,000 units.
The two companies will also enhance their technical collaboration. Naver's capabilities in building and operating large-scale GPU clusters and its expertise in hyperscale data centers will be integrated with NVIDIA's next-generation high-performance infrastructure platform, DSX.
This partnership aims to maximize data center operational efficiency and significantly improve business viability. Additionally, they will begin next-generation technological cooperation in spatial intelligence, utilizing Naver's own spatial modeling and street view data to build the 'Seoul World Model' based on NVIDIA's world foundation model, Cosmos.
Recently, Naver joined the 'NVIDIA NeMoTron Alliance,' which includes 12 top-tier global AI companies such as Cursor, Mistral AI, and Perplexity. Naver aims to accelerate the performance enhancement and global applicability of HyperCLOVA X by combining its own data and accumulated learning expertise with the collaborative technological developments from NeMoTron.
Chairman Lee Hae-jin expressed optimism about the alliance, stating, "This partnership will provide concrete alternatives for regions and countries around the world to build their own sovereign AI capabilities. It signifies a pivotal moment for Naver's technological infrastructure competitiveness to advance into the global market."
* This article has been translated by AI.
Copyright ⓒ Aju Press All rights reserved.