Seoul: South Korea's Naver Corp. announced its collaboration with Nvidia Corp. to establish a large-scale, global artificial intelligence (AI) infrastructure, beginning with the expansion of its data center in Sejong.

According to Yonhap News Agency, Naver will employ Nvidia's DSX platform to augment its Gak Sejong data center in Sejong, central Korea. The initial phase targets an expansion to 55 megawatts by the first half of 2027, with plans to extend its capacity globally, aiming for a gigawatt scale. One gigawatt represents nearly four times the current maximum data capacity of Naver's Gak Sejong center and can support hundreds of thousands of Nvidia's latest graphic processing units (GPU).

Additionally, Naver intends to join Nvidia's Nemotron Coalition, a worldwide collaboration among open model builders and AI developers, to enhance its large language model, HyperCLOVA X. The partnership will also explore physical AI, with Naver developing a "Seoul World Model." This project will integrate its urban street-view data and spatial modeling technology with Nvidia's Cosmos world foundation models.

Naver highlighted that the partnership with Nvidia extends beyond technical aspects to encompass the entire value chain, from recognizing global demand to capital collaboration. The two companies will share both the risks and benefits stemming from this collaboration.

The announcement coincides with Nvidia CEO Jensen Huang's visit to Seoul for a four-day series of meetings with leaders of major South Korean conglomerates and researchers. These discussions may lead to further cooperation in AI and robotics. Huang is set to visit Naver's 1784 building in Gyeonggi Province, south of Seoul, which serves as a testing ground for Naver's technological innovations, including cloud, AI, and robotics technologies.