Naver is stepping up to become a global artificial intelligence (AI) infrastructure company in partnership with Nvidia. The strategy is to join the global "AI alliance" led by Nvidia to rapidly upgrade its infrastructure and technological capabilities in one move, securing a comparative advantage in its competition with Big Tech players. The two companies also unveiled a plan to build a mega-scale AI cloud to preempt the surging global demand for AI computing and to create "Asia's CoreWeave."
Nvidia CEO Jensen Huang made the remarks at a media scrum held Monday at Naver's headquarters, 1784. "Nvidia has chosen Naver and plans to expand AI factories not only in Korea but across Asia," he said. Explaining the background of the collaboration with Naver, Huang said, "Naver has developed world-class AI technology and cloud," adding, "We co-developed Korea's first AI model with Naver, which was also our first supercomputer customer and partner in the Southeast Asian region."
Through this collaboration, the two companies agreed to jointly build a mega-scale global AI factory at the gigawatt (GW) level. The plan is to build "sovereign AI" infrastructure in countries around the world based on Nvidia's AI factory optimization platform, the "DSX architecture." As AI rapidly spreads worldwide, AI computing demand, previously concentrated in the United States and China, is rapidly growing in Southeast Asia as well as Europe and the Middle East.
Naver and Nvidia plan to first secure leadership in the AI infrastructure ecosystem in the Asian market, which holds high growth potential. Through this, they plan to secure influence comparable to CoreWeave, the leader in the AI cloud field. Naver expects to generate 20 trillion won in revenue from its AI factory business through this effort, roughly half of its total earnings.
"Starting with the joint construction of a 200-megawatt (MW)-class AI factory with Naver, we plan to expand to a 1-gigawatt (GW) scale in the future," Huang said. One GW is about four times the maximum capacity of Naver's data center "Gak Sejong." He added, "Once the AI factory is built, Naver will become a company that grows 10 times more than it is now."
The two companies will also strengthen cooperation in areas such as Nemotron and physical AI. Nemotron is Nvidia's open AI foundation model, and the "Nvidia Nemotron Alliance" includes 12 global AI companies such as Mistral AI and Perplexity. Naver became the first Korean company to join it.
"Nvidia's frontier AI lab is forming a partnership with Naver," Huang said. "We are teaming up in the Nemotron Alliance to build an open frontier AI model with Naver." Naver plans to use Nemotron technology to upgrade the performance of its hyperscale AI model, HyperCLOVA X.
Huang also stressed the need for cooperation with Naver in the physical AI field. "Naver has built robotics systems for a long time, and just a moment ago when I was upstairs, I enjoyed an iced coffee served by a robot. To further accelerate this, we will collaborate with Naver," he said. "The two companies will build a partnership in artificial intelligence and robotics and lead the world."
As market expectations grew over the strengthened cooperation with Nvidia, Naver's shares rose 9.2% to close higher, even as the KOSPI plunged more than 8% the previous day.
In his final greeting that day, Huang shouted "Go Naver! Go Korea!", lifting the mood at the venue with his characteristic cheer. Expressing a sense of familiarity, he said, "When I come to Korea in the future, please call me 'K-Jensen,'" and Naver Chairman Lee Hae-jin responded, "As a token of gratitude for promoting Korea widely abroad, I will pay for samgyeopsal (grilled pork belly) with CEO Huang for life."
