The Korea Institute of Science and Technology (KIST) has officially launched an industry-academia-research consortium to lead the standardization and joint demonstration of thermal management technology for artificial intelligence data centers (AIDC) in Korea. The initiative aims to grow the domestic technology ecosystem, as securing next-generation thermal management technology has emerged as a key challenge for AI industry competitiveness.
According to the science and technology community on the 24th, KIST held the "2026 AI Data Center Thermal Management Council Founding and Commemorative Symposium" at its main campus in Seongbuk-gu, Seoul, the previous day, announcing the launch of the industry-academia-research cooperation framework.
The event, co-hosted by KIST and Seoul National University's Immersion² research group, drew about 300 experts from more than 45 industry-academia-research institutions, including industry players such as Samsung Electronics, SK hynix, and LG Electronics, as well as Seoul National University, Sungkyunkwan University, Yonsei University, and the Electronics and Telecommunications Research Institute (ETRI).
The launch of the consortium comes as the spread of generative AI and hyperscale AI services has increased heat generation from high-performance AI chips, raising the need to develop source technologies in next-generation cooling technology, high heat-dissipation materials, and semiconductor packaging.
The "AI Data Center Thermal Management Council" has set as its main goals the standardization of high-reliability, high-efficiency thermal management technology for AI data centers, the establishment of a technology roadmap, the preparation of testing and reliability standards, and the promotion of joint demonstration projects. It also plans to strengthen the technological competitiveness of Korea's AI data center industry through technology exchange and joint standard development in the fields of materials, devices, packaging, cooling, and systems.
At the commemorative symposium held that day, the latest technologies and research trends in AI data center thermal management were presented, focusing on server and rack, as well as chip and package thermal management, including two-phase cooling and cold plates, immersion cooling, refrigerants for AI data centers, and heat-dissipation solutions for high-performance AI chips.
Participating institutions also discussed the council's operational direction and mid- to long-term development strategy, and agreed to cooperate on drafting standards and technology roadmaps in the thermal management field, establishing testing and reliability standards, and promoting joint demonstration projects.
Jang Jun-yeon, vice president of KIST, said, "As AI chip performance improves rapidly, the importance of thermal management technology to operate it stably is also growing." He added, "We expect this council to connect domestic capabilities from materials to packaging and cooling systems into one, and to become a focal point of cooperation that enhances the competitiveness of Korea's AI data center industry."
