Naver, Nvidia Team Up on Gigawatt-Scale AI Factory

55MW launch set for 2027, with expansion to gigawatt scale

Partnership combines Nvidia’s DSX platform with Naver technology

Alliance targets AI infrastructure leadership in Asia, the Middle East and Europe

Naver said it has formed an integrated partnership with Nvidia to build a gigawatt-scale global AI factory, extending beyond a technology alliance to cover the full value chain from demand development to capital cooperation.

Naver Chairman Lee Hae-jin and Nvidia Chief Executive Officer Jensen Huang met on June 8 at Naver’s 1784 headquarters to discuss a detailed business roadmap and specific cooperation measures for a joint push into global markets, the company said. The companies have already agreed to work together to secure leadership in the AI infrastructure ecosystem across Europe and the Middle East as well as the Asia-Pacific region.

Under the partnership, Naver will participate as a core global partner that shares responsibility for both performance and risk. The company plans to start operations with a 55-megawatt AI factory in 2027, expand to 100 megawatts later that year and then to 200 megawatts in 2028, with the ultimate goal of building a gigawatt-scale facility. At 1 gigawatt, the facility would be about four times the maximum capacity of Gak Sejong, Naver’s largest hyperscale data center in South Korea, and would be able to accommodate hundreds of thousands of Nvidia’s latest GPUs simultaneously.

The companies also plan to expand their technology cooperation across multiple areas. Naver’s capabilities in building and operating large in-house GPU clusters, along with its hyperscale data center expertise, will be combined with Nvidia’s next-generation high-performance infrastructure platform, DSX. The aim is to maximize data center operating efficiency and improve project economics.

The partnership will also extend into spatial intelligence. The companies plan to build a Seoul World Model by applying Naver’s in-house spatial modeling and street-view data to Cosmos, Nvidia’s world foundation model. Naver recently joined the Nvidia Nemotron Alliance, a group of 12 global AI companies including Cursor, Mistral AI and Perplexity, becoming the first South Korean company to participate. It plans to combine the alliance’s joint technology development with its own data and training expertise to upgrade HyperCLOVA X and broaden its global applicability.

“This alliance is highly encouraging because it gives us a concrete alternative that can help regions and countries around the world build their own sovereign AI capabilities,” Lee said. “The partnership is also meaningful because it creates an opportunity for Naver’s technology infrastructure competitiveness to take another step forward in the global market.”

Hong Min-seong, Hankyung.com reporter mshong@hankyung.com