[비즈한국] Nvidia CEO Jensen Huang personally visited Naver035420's second headquarters, '1784,' in Bundang, Seongnam-si, Gyeonggi-do, on the afternoon of the 8th. This comes three days after his so-called 'pork belly meeting' with Naver Global Investment Officer (GIO) Lee Hae-jin near Hongdae in Mapo-gu, Seoul, on the 5th. The visit took place immediately after the two companies announced a joint plan to build a gigawatt (GW)-scale, ultra-large global AI factory, and the blueprints for their future cooperation were unveiled on-site that day.
Blueprints for an AI Infrastructure Alliance Unveiled
“A 200MW (megawatt)-scale AI factory is a supercomputer of extreme scale, but this is just the beginning. We will expand it to 1 gigawatt (GW) in the future, and once the data center is completed, Naver will become a company 10 times larger than it is now.”
During the media briefing that day, CEO Huang presented the direction of the cooperation with Naver along three pillars: frontier AI models, AI cloud/AI factories, and robotics.
First, the two companies will expand cooperation in frontier AI model development, centering on Nvidia's open AI model ecosystem, the 'Nemotron Alliance.' Naver recently became the first domestic company to join the Nemotron Alliance, which includes global AI players such as Cursor, Mistral AI, and Perplexity.
CEO Huang praised Naver, stating, “The Naver AI team possesses world-class expertise,” and added, “Naver is a world-class AI developer.” He continued, “General-purpose AI alone cannot meet all the needs of nations, enterprises, and industries. Based on an open frontier model, Naver will be able to develop it in a way that fits cloud, service, and robot sectors.”
The second pillar is the construction of AI clouds and AI factories. The two companies will jointly pursue an ultra-large AI infrastructure project to respond to the growing global demand for AI. According to Naver, the plan is to begin operation with 55MW in the first half of 2027, expand to 100MW in the same year and 200MW by 2028, and eventually build a GW-scale AI factory in the long term.
1GW is about four times the maximum capacity of 'Gak Sejong,' Naver's largest hyperscale data center in Korea, and is enough to accommodate hundreds of thousands of Nvidia's latest GPUs simultaneously. The two companies also agreed to jointly enter markets in Europe and the Middle East, beyond the Asia-Pacific region. GIO Lee emphasized, “Naver was the first company in Asia to introduce Nvidia GPUs to build a SuperPOD, and we have stably operated services on the largest scale in Asia. We are not a company that says we will build a cloud or AI factory from scratch; we are a company that is already prepared.”
He continued, “Because we have built and operated data centers ourselves, I believe Naver is the only company that can immediately satisfy the rapidly increasing demand for GPUs and AI,” adding, “If we expand our AI factories not only in Korea but across Asia, it will be a huge opportunity for Naver.”
This cooperation goes beyond a simple technical supply relationship. Naver intends to participate as a core global partner in the integrated partnership program newly promoted by Nvidia, participating in the entire business value chain from discovering global demand to capital cooperation. The structure is designed to share both business achievements and risks.
CEO Huang also mentioned his partnership with Korean companies, including Naver, during his GTC Taipei keynote speech held in Taipei, Taiwan, on the 1st.
The third field of cooperation is robotics. After touring the 1784 building, CEO Huang evaluated it by saying, “I saw a robot delivering iced coffee. This place is the company of the future.” He added, “Naver has been building robotic systems for over 10 years, and Nvidia will further accelerate the development of robotic technology alongside Naver.” The building is considered a fusion of Naver's key technologies, including robots, autonomous driving, AI, and cloud.
The two companies will also expand cooperation in the field of data center operation. Naver will combine its accumulated experience in building and operating large-scale GPU clusters and its hyperscale data center know-how with DSX, Nvidia's next-generation infrastructure platform. They also plan to pursue cooperation in the field of spatial intelligence, such as building a 'Seoul World Model' using Naver's spatial modeling technology and street-view data for Nvidia's world foundation model, 'Cosmos.'
Jensen Huang Praises Naver, Lee Hae-jin Shows Confidence
Regarding Korea's industrial competitiveness, CEO Huang said, “Korea is a very rare country with world-class competitiveness in manufacturing, heavy industry, electronics, and software simultaneously. Naver also possesses world-class capabilities in AI and robotics, and is now in a position to play an important role in the global AI cloud market.”
Prior to the media Q&A, the first on-site event after CEO Huang's arrival was a webtoon event featuring the creators of the Naver Webtoon 'The Greatest Estate Developer.' In a scenario where a young person wanting to achieve both work and happiness asks for advice, CEO Huang wrote, “Don’t WORRY! I HAVE GPUs!”
The two then appeared together on a live broadcast on Naver's streaming platform, Chzzk. In the live broadcast, CEO Huang said, “Korea is the country that invented e-sports. The partnership with Naver has been a very valuable asset for a long time, and today we have further expanded that relationship.” The broadcast recorded approximately 60,000 concurrent viewers.
GIO Lee Hae-jin, Naver CEO Choi Soo-yeon, and Naver Cloud CEO Kim Yu-won were present at the site. The event venue on the first floor of the office building drew high interest, with reporters, Naver employees, and local residents gathering in crowds.