The Institute of Information & Communications Technology Planning & Evaluation (IITP) is strengthening its artificial intelligence (AI) research and development (R&D) functions as the era of agentic AI and physical AI arrives. The institute aims to serve as a catalyst across various AI fields, just as it nurtured neural processing unit (NPU) companies that have emerged as key AI chips.
"As agentic AI and physical AI are actually applied to industrial sites and services, the productivity revolution of the AX 2.0 era is beginning," IITP President Hong Jin-bae said at the "2026 Performance Media Day" held at Post Tower in Mapo-gu, Seoul, on Monday. "We will become a powerhouse for AI R&D and talent development that contributes to Korea becoming a top-3 AI power."
IITP is an R&D specialized institution under the Ministry of Science and ICT (MSIT). This year, it is supporting projects totaling 1.8996 trillion won, including 1.137 trillion won for technology development and 574 billion won for talent development. Hong cited domestic NPUs as a representative achievement. Preemptive AI chip R&D investment that began in 2020 became the foundation for cultivating four domestic NPU companies, including FuriosaAI, Rebellions, DeepX, and Mobilint. "Excluding the United States, there are not many countries with NPU package capabilities spanning both server and edge domains," Hong said. "Now that the AI market is shifting toward inference, Korea has laid the foundation to challenge the global market."
IITP also presented a blueprint to nurture defense AI companies that could grow into a "Korean Palantir." To this end, it is pursuing the establishment of a system that constantly connects private-sector AI capabilities with defense demand by building military-industry-academia cooperation centers at five major hubs nationwide, including Daejeon, Yongsan, Yangjae, Busan, and Pangyo. In the physical AI field, it will begin developing a world model this year.
"Securing 'full-stack' capabilities that integrate core technologies such as AI models, AI chips, networks, and security is more important than anything else," Hong stressed. "We will further strengthen the performance-oriented R&D support system so that the AI and information and communications technology (ICT) R&D achievements accumulated so far can lead to tangible value for industrial growth and people's lives."
