South Korea expands AI infrastructure with Nvidia GPUs and public service plans

South Korea is advancing its artificial intelligence infrastructure through significant hardware investments and a push for widespread public access. The Ministry of Science and ICT plans to launch a ChatGPT-like chatbot by the end of the year, supported by an allocation of 256 Nvidia B200 GPUs. This initiative aims to provide free, unlimited AI services to citizens, with a long-term vision to provide an AI agent for every citizen by 2027.

To foster a competitive domestic ecosystem, the government has mandated that 30 percent of AI model usage within the service must be allocated to domestically developed models. While the government estimates that 44.5 percent of the population already uses generative AI, industry officials have raised questions regarding the long-term financial sustainability of the project, specifically concerning how token costs will be covered.

In the non-profit sector, a significant gap exists between individual and organizational adoption. While 92.7 percent of Korean non-profit workers utilize generative AI in their personal workflows, only 26.8 percent of organizations have adopted the technology at an institutional level. Experts suggest that while individuals are rapidly learning to use these tools, organizations lack the necessary infrastructure, data management policies, and long-term capacity to sustain AI integration.