Nvidia CEO Jensen Huang called Naver a "global AI and cloud corporations" and said the two sides would expand their collaboration into AI models, cloud, and robots. Naver founder and board chair Lee Hae-jin said, "Naver is not a company that says it will build an AI factory in the future; it is already prepared," adding that the partnership with Nvidia will be an opportunity to grow Naver's global AI infrastructure business.
Huang and the chair spoke at a media event at Naver's 1784 headquarters in Seongnam, Gyeonggi, on the afternoon of the 8th, where they explained why Naver became a key partner for Nvidia and the future direction of their cooperation after the global AI factory joint build plan released that morning.
Huang called Korea a "very special country." He said, "Korea's natural resource starts with its people," and "Korea has remarkable expertise in manufacturing, heavy industry, electronics, and software." He added, "It is very rare for one country to be world-class in all these areas." Because AI infrastructure simultaneously requires semiconductors, power, data centers, software, and robot operations capabilities, this is interpreted to mean that Korea can have strengths in the next-generation AI industry.
Jensen Huang stressed that Naver is a global AI corporations and a cloud corporations. Huang said, "Naver is located in a relatively small country, but it has developed world-class AI and cloud technologies," calling it "an amazing achievement."
◇ "A 200MW AI factory is only the beginning"… three pillars of cooperation: models, cloud, robots
Huang laid out three pillars of cooperation with Naver. First is developing open frontier AI models through the "Nemotron Alliance" led by Nvidia. He said, "Because of Naver's world-class AI capabilities, we were able to welcome Naver as a partner in the Nemotron Alliance," adding, "Together we will advance open frontier AI models."
The Nemotron Alliance is a model collaboration that Nvidia is pursuing with global AI corporations. Naver joined the alliance—participated in by 12 global AI corporations including Cohere, Mistral AI, and Perplexity—as the first domestic corporations. Based on this, Naver plans to enhance the performance of its own large-scale AI model HyperCLOVA X and secure global general-purpose usability.
Second is AI cloud and AI factories. Huang said, "Naver is already a world-class cloud service provider in Korea and is expanding overseas," adding, "We will build a 200MW AI factory with Naver." He continued, "Two hundred megawatts is an extremely massive supercomputer, but this is only the beginning."
Naver and Nvidia released on the morning of the 8th that they would jointly pursue building gigawatt-scale, ultra-large AI factories in Asia, the Middle East, and Europe. With the hyperscale data center "Gak Sejong" as a core base, Naver plans to first run 55MW in the first half of 2027, then expand to 100MW that year and 200MW in 2028. The long-term goal is a gigawatt-scale AI factory. One gigawatt can simultaneously accommodate hundreds of thousands of Nvidia's latest GPUs.
Third is robots and physical AI. Huang said, "The next wave of AI is robots," adding, "Korea has great strengths in this area." He said, "Naver has advanced robot technology for more than 10 years," adding, "Today upstairs, I even drank an iced coffee delivered by a robot." He cited the robot system operating at the Naver 1784 headquarters as an example of future AI services.
◇ Lee Hae-jin: "Naver is almost the only company that can handle AI factory demand"
The chair cited early GPU investment, large-scale service operation experience, and in-house data center capabilities as reasons Naver became Nvidia's AI factory partner. He said, "Naver was the corporations that first adopted Nvidia GPUs and built a SuperPOD," adding, "We have invested from early on."
He added, "Naver is a company that stably operates the largest service in Asia," and "It is not that we will build a cloud and an AI factory in the future; we are already very well prepared." He also said, "We build and operate our own data centers well," adding, "In the AI market where demand is rising rapidly now, we believe we are the only company that can meet that demand."
The chair explained that Nvidia's intent to expand AI factories with Naver not only in Korea but also into the Asian market is due to these prepared capabilities. He said, "In that respect, Nvidia chose us and is trying to expand AI factories with us not only in Korea but also into Asia." He added, "A gigawatt-scale AI factory is truly enormous," and "If we can do this together, it will be a really big opportunity for Naver."
Huang said the direction of cooperation between Naver and other Korean corporations differs in each case. He said, "Each company has its own expertise, and we work with each company in its own way," adding, "Naver is already Korea's global cloud service provider and is now expanding worldwide." He continued, "The world now needs AI cloud," adding, "I am very pleased to build Naver AI Cloud around the world with Chair Lee Hae-jin and the Naver team."
◇ "General-purpose AI is not enough"… targeting AI specialized for Korean, manufacturing, and robots
Huang cited demand for country- and industry-specific AI as the reason model collaboration with Naver is necessary. He said, "Frontier AI labs like OpenAI and Anthropic are doing great work, and we all use those technologies," but added, "For many countries, corporations, and specialized domains, general-purpose intelligence alone is not enough."
He went on, "AI needs to be adapted to the Korean language and Korean culture, and in fields like science, digital biology, robots, and manufacturing, its functions must be enhanced for specific purposes." This means that beyond using generic large language models (LLMs) as-is, AI needs to be tailored to each country's language and industrial settings.
Huang said, "Naver has a great deal of expertise in AI model development," adding, "By combining our training data and expertise, we can create a great starting point for frontier models, and Naver can tune and develop them for its services and research areas."
The two companies will also expand infrastructure technology cooperation. They plan to combine Naver's accumulated capabilities in building and operating large-scale, in-house GPU clusters and hyperscale data center know-how with Nvidia's next-generation high-performance infrastructure platform, DSX, to improve data center operating efficiency and business viability. They will also pursue spatial intelligence cooperation to build a "Seoul World Model" by leveraging Naver's spatial modeling and Street View data in Nvidia's world foundation model, Cosmos.
The event also served as a venue to showcase the two companies' technology alliance. The chair said, "In Korea, I think this is the first time I have seen such an atmosphere in which an overseas businessperson is so loved, respected, and welcomed by so many people," adding, "Beyond Nvidia's technological prowess and market cap, it seems that Jensen Huang's own character and personality are what make so many people love him."
Huang said, "During this visit to Korea, I received a new name at Seoul National University," adding, "My new name is 'K-Jensen.'"