While China races ahead in Physical AI, including Humanoid Robot and Autonomous Driving mobility, by amassing vast real-world data, Korea still appears to rely on limited data infrastructure.
According to the report "AI and Humanoid Robot industry led by China," released on the 9th by the National Information Society Agency (NIA), Chinese robot corporations are strengthening the global competitiveness of humanoid technology by focusing on data-driven learning and model advancement.
In particular, the Start - Up AgeeBot, backed by Tencent, set up a "robot data factory" in Shanghai in 2023 and is accumulating 30,000 to 50,000 real-world data entries per day by deploying about 100 robots and 200 personnel. The high-quality data generated at this factory is said to account for 80% of the real-world data used to train Nvidia's Humanoid Robot foundation model "Isaac GR00T N1."
The Chinese government is moving to secure empirical data using hospitals, schools, and public facilities across nine industries, including manufacturing, education, services, health care, and property management. The report analyzed, "China is pushing a 'new industrial revolution' in which robots perform production tasks in actual factories."
China is dominant not only in data but also in the talent pool. Last year, the number of students majoring in robotics at Chinese universities surpassed 580,000, accounting for 42% of the global total.
By contrast, Korea's acquisition of Physical AI data remains at a basic stage. Another NIA report, "Global trends and response strategies for Physical AI," noted that while the government-run AI Hub serves as the foundation for domestic Robotics and Autonomous Driving research, it faces major limits in environmental diversity and international compatibility and lacks real-time data.
The report said, "Deep datasets such as robot behavior data and sensor and environment data must be built and opened step by step," and suggested, "Verification infrastructure linked to Digital Twin environments at key hubs such as ports, airports, hospitals, and smart buildings is needed."
It also stressed that "to secure interoperability among sensors, robots, and AI platforms, the government should form a standardization consortium and manage data quality, safety, and compatibility in an integrated manner."