GIST to build national AI research platform with W75b over 8 years

South Korea is moving to build a university-centered AI research ecosystem aimed at tackling hard scientific and technological challenges and strengthening the country's future research competitiveness.

The Gwangju Institute of Science and Technology (GIST) announced it has been selected for the "central hub infrastructure" category of the Ministry of Science and ICT and National Research Foundation's AI-Based University Science and Technology Innovation Project.

Under the project, GIST will receive 75 billion won ($50.3 million) over eight years through 2034 to build and operate a shared AI research infrastructure for universities.

AI has increasingly become a core tool in scientific research, with applications in drug development, new materials discovery, semiconductor design and climate forecasting.

Major economies including the United States, Europe and Japan have made the convergence of AI and scientific research a national strategy, investing in dedicated research infrastructure. South Korea, by contrast, has AI computing resources, data and research environments scattered across individual institutions, leaving the country without a national-level shared research foundation that researchers can use collectively.

The government is pursuing the "K-Moonshot" initiative to keep pace with global AI competition, and plans to use this project to build a nationally shared AI research infrastructure and foster a sustainable, university-centered ecosystem for science, technology and AI convergence research.

GIST plans to build a shared high-performance computing and AI infrastructure (HPC-AI) that meets the demands of scientific research, and will operate an integrated research platform called K-ImagineRI (Korea-Imagine Research Infrastructure) to give researchers easy access to these resources.

The platform will integrate ultra-high-performance GPU-based AI research infrastructure, research data and AI model sharing, research workflow support, and joint research support across research centers — enabling researchers to carry out the entire research and development cycle, from planning and data analysis to AI model development, application and results sharing, all within a single platform.

The project is significant in that it goes beyond simply expanding GPU hardware, aiming instead to build a comprehensive AI research foundation encompassing high-performance computing resources, data and model sharing, research collaboration, technical support and education.

The initiative is expected to reduce the time and cost researchers spend setting up individual research environments, while enabling shared use of data and AI models accumulated during the research process — accelerating the era of AI-driven science and technology innovation (AI4S&T).

"The success of AI4S&T hinges on how easily researchers can access the computing resources and data they need," said Kim Jong-won, dean of the GIST College of AI. "GIST will build a researcher-centered, AI-based shared infrastructure and contribute to creating a new innovation ecosystem for science and technology research in Korea."