China's data regulator is set to develop standards for embodied AI data, a move that could significantly impact the industry's growth. The National Data Administration announced it will guide local authorities on the work and support companies investing in data resources, just ten days after seven companies requested public data infrastructure for embodied AI and common data standards.
The development of standards for embodied AI data is crucial, as the industry is becoming increasingly data-driven. Research institutes, technology companies, and humanoid robotics organizations attended a meeting on September 10, chaired by Liu Liehong, head of the agency. The meeting highlighted the need for standards to support the industry's growth.
Embodied AI foundation models require massive amounts of real training data, estimated to be around 10 million hours. However, only between 100,000 and 1 million hours of high-quality data exist worldwide. China is addressing this shortage by building a supply of embodied AI training grounds, with over 70 already operating across more than half of the country's provincial regions.
In contrast, Europe has written rules for this data without collecting much of it. The Data Act, which came into effect on September 12, 2025, governs who can access the data generated by connected products, including industrial machinery. However, the European Union has not set up data labs as promised in its Data Union Strategy, leaving companies like NEURA Robotics to fill the gap by building their own training grounds.
The development of standards for embodied AI data in China is expected to have a significant impact on the industry, but it does not guarantee the country's dominance in the market. With only 23% of Chinese enterprises satisfied with the robots on offer, there is still room for growth and improvement.
This article was written with the assistance of AI.
News Factory APP - agentic news to boost your SEO & AEO.