Korea's shipbuilding and marine industrial complexes are transforming into AI transformation (AX) hubs centered on Software-as-a-Service (SaaS)-based collaboration platforms.
Busan's Myeongji-Noksan Industrial Complex will optimize quality prediction and equipment anomaly detection through SaaS solutions. The complex will maximize operational efficiency by analyzing and predicting logistics, energy, and traffic flows in real time using a virtual factory, also known as a digital twin. The virtual factory technology recreates physical spaces in virtual environments, where AI analyzes data from the complex in real time to predict problems and verify the effects of policy and facility investments.
The Ministry of Trade, Industry and Energy (MOTIE) and Korea Industrial Complex Corporation (KICOX) have launched an industrial complex AX support project, accelerating manufacturing innovation. The government plans to establish AI-based operating systems across all industrial complexes by 2028 through shared infrastructure and standardized data construction. Industrial complexes are considered advantageous for AX activation because companies within the value chain are concentrated in a single location. With 97% of tenant companies being small and medium-sized enterprises with fewer than 50 employees, government-level AX support is expected to enhance policy effectiveness.
According to KICOX on May 20, ten domestic industrial complexes selected through the AX support project have begun building AX demonstration complexes.
The ten industrial complexes are Banwol-Sihwa in Gyeonggi Province, Changwon in South Gyeongsang Province, Myeongji-Noksan in Busan, Mipo in Ulsan, Cheonan in South Chungcheong Province, Gunsan in North Jeolla Province, Cheomdan in Gwangju, Yeosu-Daebul in South Jeolla Province, and Hupyeong in Gangwon Province.
The project focuses on creating AI operating systems at the industrial complex level rather than for individual companies. The approach involves replicating and spreading standard models such as AX leading factories and virtual factories to multiple tenant companies.
AX leading factories serve as best-practice examples that substantially improve on-site KPIs such as defect rates, uptime, and quality. On-site KPIs are indicators that quantitatively evaluate changes in productivity, quality, safety, and costs at manufacturing sites following AI adoption. Virtual factories allow AI to pre-verify design and processes, reducing failure costs and lead times while enhancing overall productivity and efficiency across the complex.
The ten selected complexes will establish AX leading cases by value chain through 2028. They plan to develop modularized systems that can be applied to multiple companies. Ulsan Mipo and Yeosu will develop petrochemical-specialized modules based on process, facility, and safety data from oil refining and petrochemicals. Changwon in South Gyeongsang Province and Cheomdan in Gwangju will standardize the process of "benchmarking → replication to first and second-tier suppliers and SaaS adoption" centered on large corporation-led AX leading factories. SME manufacturing, specialized processes, and biotechnology sectors will maximize cluster effects with bio labs. Banwol-Sihwa, a foundry and parts materials complex, will combine injection molding (product inspection) and printed circuit board (PCB)-centered standard models with open labs at Hanyang University and Korea University of Technology and Education to create modules immediately available to SMEs. Cheonan, with its high proportion of display industry companies, will design AX leading factories to immediately demonstrate defect reduction and efficiency improvements, then convert these into SaaS for expansion to parts and equipment suppliers. Hupyeong in Gangwon Province, where biotechnology companies are concentrated, aims for one-stop verification and system expansion by integrating open labs, leading factories, and virtual factories into a Wise Factory combining AI and Internet of Things (IoT).
"When indicators such as defects, downtime, and energy consumption improve at leading factories, the AI packages and application guides for those processes will immediately transfer to other companies in the complex," a KICOX official said. "If virtual factories reduce failure costs and open labs continue to discover promising new companies, industrial complexes will develop a self-upgrading structure centered on shared data and model repositories and training systems."
