The government invested 1.4131 trillion won over 5 years... Expanding corporate cooperation
(Seoul = Yonhap News) Reporter Kwon Ha-young = With a global artificial intelligence (AI) evaluation agency recognizing Korea as a 'firm third' in the AI competition, Korea's attention is now turning to the next battlefield.
As the stage of AI competition moves beyond the digital space and into the real world, a solid manufacturing base in the semiconductor, shipbuilding, and defense industries is attracting attention as a new competitive asset in the physical AI era.
◇ Korea ranked 3rd in AI, the next task is the ‘real world’
Recently, Artificial Analysis, a global AI evaluation agency, evaluated Korea as 'clear #3' in the global AI competition, following the United States and China, and delivered a certificate commemorating this to the Ministry of Science and ICT.
The next task for Korea, which has proven its AI model technology competitiveness, is to connect this capability to actual industrial competitiveness. This is because beyond AI that understands and creates text and images, physical AI that recognizes the real world and connects it to the actual behavior of robots and facilities is emerging as a front line in new technological competition.
In particular, there is analysis that Korea can occupy an advantageous position in this trend. This is because it has a wide range of industrial sites where physical AI can be actually applied and verified, including semiconductors, automobiles, shipbuilding, steel, batteries, and defense industries.
An official in the domestic AI industry said, "Combining AI and manufacturing is not a matter of choice, but a matter of survival. AI that deals with text or images can be learned anywhere as long as there is data, but physical AI can only be properly learned by learning the noise, vibration, and unpredictable situations of an actual factory, and Korea's manufacturing site itself can become an AI training ground."
◇ Government invests 1.4 trillion won in physical AI… Demonstration at manufacturing site
The government is elevating this opportunity to a national level strategy. Physical AI was designated as one of ‘Korea’s three major leap forward megaprojects’ along with semiconductors and AI data centers.
In the 'Strategy to Secure Core Competitiveness in Physical AI' announced last July, data, technology, diffusion, and ecosystem were set as four major axes. The goal is to build an ecosystem that spans from data acquisition and core technology development to field verification and commercialization.
The key is to secure full-stack competitiveness by securing actual and synthetic data from industrial sites, developing core technologies such as world models and physical AI foundation models, and connecting them with hardware such as robots, sensors, and AI semiconductors.
AI transformation in the manufacturing industry is also a major stage of this strategy.
The Ministry of Trade, Industry and Energy is promoting 'M.AX (Manufacturing AI Transformation)', which refers to the transformation of manufacturing AI, and is expanding the ecosystem that applies AI to overall manufacturing, including AI factories, AI robots, and AI semiconductors.
With the goal of building 500 AI factories by 2030, the plan is to support high-quality manufacturing data collection and AI model development and verification, and to expand the scope of AI application from the production process to production planning, supply chain, and inventory management.
Large-scale national projects also began in earnest.
The government will invest a total of 1.4131 trillion won over five years from 2026 to 2030 to develop autonomous factory operation technology in Jeonbuk and precision control technology for manufacturing processes in Gyeongnam and demonstrate them in actual industrial sites. It is a structure that directly utilizes the region's major industries as a basis for physical AI development and verification.
◇ Expanding corporate cooperation… From shipbuilding welding to defense simulators
In line with the government's strategy, companies are also rapidly expanding their attempts to connect each industrial site with AI technology.
NC[036570] AI is simultaneously participating in national projects in Jeonbuk and Gyeongnam, expanding the scope of physical AI application to various industrial sites such as steel, shipbuilding, and national defense.
Jeonbuk is developing technology to implement factory equipment and workspaces in a 3D virtual environment and generate AI learning data so that equipment and robots can autonomously recognize and respond to the entire factory situation.
In Gyeongnam, based on the world model and large-scale behavioral model (LAM), we plan to use sensors and vision data to predict on-site situations and implement precise control technology to control the behavior of facilities and robots, and apply it to shipbuilding welding and semiconductor post-processing.
We are jointly developing a robot foundation model (RFM) to be applied to industrial sites with POSCO DX. The key is to develop a VLA model that integrates vision, language, and behavior so that the robot can recognize the surrounding environment with a camera, understand verbal instructions, and then carry out actual actions.
With Hanwha Ocean, we are developing physical AI technology specialized for welding work at shipbuilding sites, and with Hyundai Rotem, we are developing a physical AI integrated simulator and robot technology that can learn and verify various battlefield situations in a virtual environment.
An NC AI official said, "For physical AI to work in the field, it must be supported not only by AI technology but also by the physical environment such as processes, facilities, and robots to implement and demonstrate it. The key is how effectively the government's national strategy, corporate technology development, and industrial field verification are aligned."
