Doosan Robotics Wins KRW 989B for On-Device AI Robots

Recap

Source: 디일렉, report of Oct. 6, 2026

According to the report, Doosan Robotics has secured two national-level research grants under the Ministry of Trade, Industry and Energy and the Korea Evaluation and Planning Institute for Technology. The total research and development budget for these two projects is approximately KRW 989 billion, with government support covering about KRW 681 billion.

The first initiative focuses on creating a "Korean-style Physical AI Collaborative Robot Integrated Platform." This involves developing next-generation collaborative robots that can perceive, judge, and control their actions autonomously without relying on external PCs or cloud infrastructure. The system will integrate domestic AI system semiconductors with neural processing units (NPUs) and real-time control functions. The company aims to produce three types of collaborative robots with payload capacities of 5kg, 10kg, and 20kg, as well as an industrial humanoid capable of dual-arm operations. Commercialization of these units is targeted for 2031.

The second project aims to automate high-precision welding tasks by combining collaborative robots with AI and digital twin technology. Doosan Robotics will demonstrate these capabilities, including tack welding and preheating, at the Doosan Enerbility Nuclear Center. The company plans to expand the application of this technology to the shipbuilding, plant, and defense sectors in the future. Park In-won, CEO of Doosan Robotics, stated that the goal is to develop physical AI collaborative robots that can learn and reproduce the skills of highly trained workers by combining on-device AI semiconductors with their own robot platform.

Context

This funding aligns with South Korea's broader strategic push to localize its AI semiconductor supply chain and reduce dependence on foreign hardware for critical industrial applications. The emphasis on "on-device" AI reflects a global trend in robotics where edge computing is preferred for its lower latency and enhanced data privacy, particularly in sensitive industries like nuclear energy and defense. Doosan Robotics has previously established itself in the collaborative robot market, and this project represents a significant step toward integrating advanced AI capabilities directly into the robot's hardware architecture.

Robot's take

The shift toward on-device AI is a critical development for industrial robotics, as it promises greater autonomy and resilience in environments where cloud connectivity may be unstable or prohibited. However, the 2031 commercialization timeline suggests that the technology is still in its early developmental stages. The integration of domestic AI chips with real-time control systems presents significant engineering challenges, particularly in ensuring the reliability and precision required for high-stakes tasks like nuclear welding. It remains to be seen whether the digital twin approach will successfully scale across diverse industrial sectors beyond the initial nuclear applications.