Seoul: The government announced plans to allocate approximately 10,000 Nvidia graphics processing units (GPUs) to small and mid-sized enterprises (SMEs), startups, and academic and state artificial intelligence (AI) projects beginning in February. This initiative is part of a broader strategy to enhance the domestic AI ecosystem, as discussed in a recent meeting chaired by Science Minister Bae Kyung-hoon.
According to Yonhap News Agency, the government invested 1.4 trillion won (US$947.2 million) to acquire the GPUs from Nvidia Corp. This marks the initial phase of a larger plan to procure 50,000 units by 2030. The GPUs will be organized into a large-scale cluster to facilitate high-speed computing necessary for AI model training and inference.
Applications for AI development projects from various sectors, including industry, academia, and research institutions, will be accepted online until January 28. Each project could receive support of up to 256 H200 GPUs or 128 B200 GPUs. Additionally, around 6,000 B200 GPUs, slated for later delivery, will be dedicated to a national effort to develop indigenous AI foundation models.
The allocation aligns with Nvidia CEO Jensen Huang's recent announcement that the company plans to deploy 260,000 GPUs to South Korea. This deployment is in partnership with the government and major corporations, such as Samsung Electronics Co., to establish extensive AI factories in the country.
Under the current plan, the government will utilize about 50,000 GPUs to construct a national sovereign AI platform. Concurrently, Samsung Electronics, SK Group, and Hyundai Motor Group are each set to receive 50,000 units, while Naver Corp. is slated to receive 60,000 units.