Naver has teamed up with Nvidia and agreed to a joint project to build a gigawatt (GW)-scale, ultra-large global AI factory. Naver said the alliance goes beyond a simple technology tie-up and is an integrated partnership that spans the entire value chain, from exploring global demand to capital cooperation. In particular, Naver will participate as a core global partner jointly responsible for the project's outcomes and risks. To capture surging worldwide AI demand, the two companies plan to move quickly to build an ultra-large AI factory that will set the standard for global AI infrastructure, starting with a 55MW project in 2027.
1GW is about four times the maximum capacity of Naver's "Gak Sejong," the largest hyperscale data center in Korea. It is large enough to accommodate hundreds of thousands of Nvidia's latest GPUs simultaneously.
Lee Hae-jin, chair of Naver's board, and Jensen Huang, Nvidia chief executive officer (CEO), will meet today at Naver's 1784 headquarters to discuss the detailed roadmap of their ongoing project and specific cooperation plans for a joint entry into global markets. The two companies have already agreed on a direction to secure leadership in the AI infrastructure ecosystem together, not only in the Asia-Pacific region but also in Europe and the Middle East.
First, Naver will use its hyperscale data center "Gak Sejong," a key hub, as a forward base. With the start of 55MW operations in the first half of 2027 as a signal, it plans to expand infrastructure overseas to 100MW within the same year and 200MW in 2028 to absorb global demand. Ultimately, this is part of a step-by-step roadmap to build GW-scale infrastructure that will dominate the market.
Furthermore, the two companies will comprehensively deepen their technological integration. Naver's uniquely accumulated capabilities in building and operating large-scale, in-house GPU clusters and its hyperscale data center know-how will be rapidly fused with Nvidia's next-generation high-performance infrastructure platform "DSX." Through this, they plan to maximize data center operational efficiency and significantly enhance business viability. They will also launch next-generation cooperation in spatial intelligence, including building a "Seoul world model" that leverages Naver's in-house spatial modeling and street view data on Nvidia's world foundation model "Cosmos."
Naver recently became the first among Korea's corporations to join the "Nvidia Nemotron Alliance," which includes 12 global top-tier AI companies such as Cursor, Mistral AI, and Perplexity. Building on this, Naver is accelerating performance enhancements and global versatility for "HyperCLOVA X" by combining Nemotron's joint technology development results with Naver's own data and accumulated training know-how.
Lee Hae-jin, chair of Naver's board, said, "It is very encouraging that this alliance enables us to present concrete alternatives for each region and country around the world to build independent sovereign AI capabilities," and added, "This cooperation is meaningful in that it provides an opportunity for the competitiveness of Naver's technology infrastructure to take a step forward into the global market."