Naver Cloud will launch a global AI factory build-out with Nvidia. By combining capabilities in AI infrastructure, models, and service operations, it aims to respond to surging demand for artificial intelligence (AI) and expand its foothold in the global AI infrastructure market.
Kim Yu-won, CEO of Naver Cloud, outlined the direction of collaboration with Nvidia at the Nvidia Cloud Partner Summit held in Taiwan on the 2nd. Kim emphasized that Naver Cloud is a full-stack technology corporations that has directly operated everything from AI infrastructure to services, and said it aligns with Nvidia's AI factory strategy spanning energy, chips, infrastructure, models, and applications.
Nvidia CEO Jensen Huang introduced Naver Cloud as a key AI-native cloud partner during the "GTC Taipei 2026" keynote on the previous day. The two companies plan to expand cooperation beyond a simple graphics processing unit (GPU) supply relationship to include AI model advancement and the physical AI domain.
Naver Cloud will use Nemotron 3 Ultra, Nvidia's open large language model (LLM), to boost HyperCLOVA X performance and will also pursue optimization of ultra-large language models and foundational technology research. On Mar. 3, it unveiled the "Seoul World Model," built with Nvidia's physical AI platform Cosmos to reproduce road environments and spatial structures across Seoul. The model trained on domestic map data and 1.2 million panoramic images collected in Seoul.
Leveraging its dedicated cloud and model operation experience, Naver Cloud will also support the build-out of sovereign AI requested by governments and corporations in each country. Sovereign AI refers to a concept in which a nation or corporations operate AI models and infrastructure tailored to their own data and regulatory environments. As Data Sovereignty and security needs grow, it is emerging as a core competitive axis in the global cloud market.
An AI factory refers to next-generation infrastructure that trains AI models and mass-produces inference services through GPU-based data centers. After the spread of Generative AI, corporations are prioritizing stable inference operations and expense efficiency over model development, accelerating cloud providers' competition to secure AI data centers.
The two companies plan to soon unveil detailed execution plans through a meeting in Korea between Naver Chair Lee Hae-jin and CEO Jensen Huang.
CEO Kim Yu-won said, "The collaboration with Nvidia is a strategic decision to expand the global AI ecosystem together beyond a supplier-customer relationship for GPUs," adding, "We will establish ourselves as a key supplier supporting Asia's AI demand and as a hub for AI infrastructure."