- MSIT presents R&D achievements at MWC 2026
- AI-integrated network management cuts power consumption by 20% and ensures stable connectivity
- Technology transfers to strengthen domestic competitiveness and reduce network operating costs
The Ministry of Science and ICT (MSIT; Deputy Prime Minister and Minister Bae Kyunghoon) announced that it showcased the next-generation network technology achievements of Korea’s industry, academia, and research institutes at Mobile World Congress (MWC) 2026, the world’s largest mobile technology exhibition held from March 2 to 5, 2026.
A total of 199 Korean companies, organizations, and research institutions participated in MWC 2026, unveiling a wide array of R&D breakthroughs and innovative services in 6G, AI, and satellite communications. Among these, the AI network technology developed by the Electronics and Telecommunications Research Institute (ETRI) and Yonsei University (Professor Seong-Lyun Kim’s Lab) supported by the MSIT’s "Broadcasting and Communication Industry Technology Development Project*" drew significant attention as a key driver for low-power, high-efficiency next-generation mobile networks.
* A bottom-up R&D project planned to reflect industry demand and technological evolution in the network sector.
1. Electronics and Telecommunications Research Institute (ETRI): Low-Power, Open 5G-Advanced Base Station Software
ETRI introduced low-power base station software technology capable of reducing the energy consumption of base station equipment—which accounts for 70% of total mobile network power usage—by more than 20%. The AI algorithms minimize power waste by predicting real-time traffic and precisely controlling resources.
This technology adopts the latest 5G-Advanced standards, the bridge between 5G and 6G. It is designed as software-based, allowing it to run flexibly across various server environments without being locked into specific hardware manufacturers, thereby increasing its industrial viability.
ETRI’s research achievement has been recognized for its excellence, securing 34 domestic and international patents and having five of its innovations incorporated into international standards; furthermore, the technology is ready for immediate commercialization upon the completion of the research. This technology is expected to be transferred to two domestic small and medium-sized enterprises (SMEs) to help secure the technological competitiveness of domestic firms and is projected to save approximately 100 billion won out of the 1 trillion won in annual communication network operational electricity costs.
Looking ahead, the research team plans to collaborate closely with domestic AI semiconductor companies to achieve a 30% reduction in energy consumption using Korean Neural Processing Units (NPUs). This initiative aims to develop reference examples for domestic AI chips and strengthen the value chain throughout both upstream and downstream industries. Furthermore, the team plans to leverage the strengths of highly versatile software-based technologies to promote their adoption across various use cases, including smart factories.
2. Yonsei University: AI-Driven Intelligent and Autonomous Antenna Control Technology
In collaboration with the Singapore University of Technology and Design (SUTD) and the U.S.-based Viavi Solutions, Yonsei University developed technology that enables AI to autonomously assess real-time conditions and efficiently control antennas*.
* Key equipment that exchanges wireless signals using beams between base stations and terminals.
High-quality AI services, such as "Physical AI," which are expected to grow exponentially in the near future, require antenna technology that can maintain stable signal transmission even in unexpected situations like high-speed movement or obstacles. The research team has integrated Large Language Model (LLM) technology, similar to that used in ChatGPT, into base station and antenna management.
This LLM-based management system allows AI to identify and predict network status in real-time to actively toggle antenna signals or determine their direction (beamforming). In experiments simulating irregular signal cycles or network anomalies, the LLM-based model demonstrated superior efficiency and stability* compared to existing reinforcement learning-based AI models.
* At identical power consumption levels: △ Irregular signal cycles: Achieved a significant jump in throughput (95.7%) and stability (29%→90%), △ Network anomalies: Showed a 3% increase in throughput and a 12.4% increase in stability.
Yonsei University’s research was exhibited at the booth of the AI-RAN Alliance*, led by global AI leaders such as NVIDIA and SoftBank.
* An international organization launched in 2024 to advance AI-RAN technology and its ecosystem.
Regarding the research achievements of ETRI and Yonsei University, a member of the telecommunications industry said, "As future network traffic is projected to increase by more than four times, power efficiency enhancement technology will contribute significantly to network stability and operation cost reduction. We expect the role of domestic low-power network technologies and equipment to grow in the process of building future low-power communication networks."
Deputy Minister Choi Woo-hyuk of the Office of Cyber Security and Network Policy at the MSIT, who visited the site that day, said, “As the global race for 6G and AI networks is intensifying, it is encouraging that the government, industry, academia, and research institutes are closely cooperating to secure the foundational technologies to lead the industry. The MSIT will continue to provide active support—from technology development to finding references and market entry—to become a first mover in 6G and AI networks”
For further information, please contact the Public Relations Division (Phone: +82-44-202-4034, E-mail: msitmedia@korea.kr) of the Ministry of Science and ICT.
Please refer to the attached PDF.