The Ministry of Science and ICT on the 23rd unveiled at KAIST a "physical artificial intelligence (AI) integrated platform" that combines sensors, control, robots, and manufacturing software needed to run factories using domestic technology, and presented a "strategy to secure core competitiveness in physical AI (draft)" based on it.
Until now, factory operation software and a substantial portion of equipment control solutions at domestic manufacturing sites have relied on foreign products. To ease this structure, the government has been promoting a "physical AI pre-verification project" centered on Jeonbuk National University and KAIST through the 2025 supplementary budget.
While the Jeonbuk National University demonstration lab is tasked with verifying flexible production AI that responds to a multi-product, small-batch, multi-process production environment, the KAIST demonstration lab focused on integrating the factory operating system with domestic technology and optimizing production schedules and logistics operations.
In particular, the KAIST demonstration lab applied core technologies needed for overall factory operations—such as sensors, controllers, robots, and AI data infrastructure—centered on domestic corporations. The government sees significance in testing the possibility of replacing factory operating systems that had relied on foreign solutions with domestic technology.
One of the core features of this platform is an operating agent known as the "AI plant manager." It uses Digital Twin-based simulations to analyze and adjust real-time logistics flows and production schedules in actual factories. The government believes this will allow small and midsize manufacturing corporations to pursue advanced factory operations without expensive foreign solutions.
The two demonstration labs are set to be operated as open test environments going forward. The government plans to link this with the main AI transformation (AX) project in the Jeonbuk region to advance autonomous factory operating systems and, in the long term, develop it into an export model in the form of a "K-manufacturing intelligent factory package."
◇ From technology development to exports: full-cycle strategy for physical AI presented
The "strategy to secure core competitiveness in physical AI (draft)" unveiled the same day focuses on building a full-cycle framework spanning technology development, demonstration, industry diffusion, and overseas expansion. Physical AI goes beyond making decisions in digital space to AI that directly perceives and acts in the real world. The government views it as a next-generation technology that can spread not only to manufacturing but also to logistics, agriculture, disaster and safety, and care and home.
The government plans to launch development of a "robot foundation model" capable of long-term, precise task execution; a "world model" that supports large-scale training data generation and virtual experiments; and a "computing platform" based on AI Semiconductor for high performance, low power, and low latency. It will also push ahead with developing core humanoid technologies, networks, and security technologies.
Second is expanding technologies for automating manufacturing processes. To that end, the plan is to widely secure equipment operation data, task behavior data, factory operation data, and virtual and synthetic data, and based on this, develop autonomous control technology for manufacturing equipment and optimization technology for factory operations. The blueprint also includes connecting this to large-scale projects applied to key industries such as automobiles, precision manufacturing, and shipbuilding, aiming for consolidation.
Along with regulatory improvements, investment attraction, and support for overseas expansion, the government plans to establish a talent development system encompassing on-site practitioners to master's and doctoral-level researchers. It also presented as concurrent tasks responding to international standardization, developing performance evaluation benchmarks, and improving safety and institutional foundations. It added that it will accelerate the industrial application of physical AI technologies by strengthening linkages among ministries as well as among associations, alliances, corporations, universities, and research institutes.
The Ministry of Science and ICT plans to incorporate industry opinions raised at the roundtable into the strategy draft and finalize it after consultations with related ministries.
Bae Kyung-hoon, Deputy Prime Minister and Minister of the Ministry of Science and ICT, said, "The next three years are an important period for securing competitiveness in physical AI," and added, "We will establish a support system that connects technology development to on-site application and industrialization." He continued, "We will also expand export models in the manufacturing sector based on the domestic factory operating systems verified at the demonstration labs."