TMTPOST--As global interest in humanoid robots continues to surge, Chinese cities are racing to position themselves as leaders in this emerging industry, even though commercial prospects remain uncertain.

In Beijing, the humanoid robotics race gained momentum with the establishment of the Beijing Humanoid Robotics Innovation Center in late 2023.

Founded with the backing of industrial giants including UBTECH, Xiaomi, and Jingcheng Mechanical and Electrical, along with Yizhuang Holding, the center was set up under the guidance of local industrial and information technology departments. Each of the first three companies holds approximately a 28.57% stake, making them the largest shareholders.

Sources revealed that these founding shareholders contributed over 600 patents related to humanoid robotics to the innovation center free of charge. Furthermore, any patents developed by the center will be made available to the shareholders at no cost.

In Shanghai, another major player emerged with the inauguration of a national-local humanoid robotics innovation center in May 2024. The center is constructing a training facility capable of hosting 100 humanoid robots for intelligent training, with the first phase set for completion in November. By 2027, the training facility aims to expand its capacity to accommodate 1,000 humanoid robots.

In March 2024, Shenzhen saw the establishment of the Guangdong Province Embodied Intelligence Robotics Innovation Center, led by the Chinese University of Hong Kong (Shenzhen) and its AI and Robotics Institute. This center is supported by various institutions and companies, including ForwardX Robotics and Leju Robotics.

Beyond the major cities, provinces such as Zhejiang, Sichuan, and Anhui are also investing in humanoid robotics innovation centers to stimulate local robotics industries.

This wave of humanoid robotics development marks a shift towards collaboration between governments and enterprises. Unlike the solo ventures of startup companies, innovation centers are leveraging public-private partnerships to drive the growth of the humanoid robotics industry by uniting companies across local industrial chains.

However, with numerous cities entering the humanoid robotics race, the question remains: who will emerge as the frontrunner?

When Tesla founder Elon Musk reignited interest in humanoid robots, China's first wave of participants were young entrepreneurs. Since 2023, several domestic humanoid robot products have been introduced, including Uisee's H1, Zhiyuan Expeditions' A1, Fourier's GR-1, Xingdong's "Little Star," and ForwardX's CL-1. Much like the early days of autonomous driving, many of these companies are led by young founders, mostly born in the 1990s.

Wang Xingxing, founder of Uisee Technology, attributed the rapid rise of the humanoid robotics industry to the advent of large AI models, which opened up new possibilities for humanoid robots to evolve towards general-purpose applications.

According to Wang, both large AI models and humanoid robotics are new technological fields, where previous experience and technology are no longer applicable, making younger generations quicker learners and more willing to dive in.

As the humanoid robotics wave continues to gain momentum, the "national teams" — humanoid robotics innovation centers — have also come to the forefront. Government-guided funds are increasingly participating in the early financing of humanoid robotics startups, a phenomenon that was notably absent during the autonomous driving boom a decade ago.

"Such deep involvement from the government in this early-stage industry is unprecedented compared to ten years ago," said Wang Guangxi, Vice President of Lenovo Group and Partner at Lenovo Capital.

He attributed this shift to changes in the capital market environment, where state-owned capital is playing an increasingly important role in emerging industries. Additionally, humanoid robotics is a complex interdisciplinary engineering challenge, spanning AI software, hardware, and application scenarios, making it difficult for any single institution to gather the necessary resources without government support.

While the establishment of innovation centers can accelerate the development of humanoid robotics, the differences in regional industrial structures have led to varying levels of supply chain efficiency.

"From a hardware supply chain perspective, the Yangtze River Delta region has a more complete supply chain," said Guo Tao, CTO of Pateo Robotics. He explained that the region has mature supply chain support for industrial robots, from basic structural components to processing and casting, and even cable connectors. Compared to industrial robots, humanoid robots have more complex components that are still in the early stages of development. Quick responses from suppliers can be critical to the success of product development.

According to business data, China's robotics-related enterprises are concentrated in cities such as Shenzhen, Guangzhou, Shanghai, and Suzhou, with each city hosting tens of thousands of companies. This clustering of robotics companies lays a solid foundation for the supply of humanoid robot components.

Furthermore, the rapid development of the electric vehicle industry in China has provided strong support for humanoid robotics. The advancements in electric motors, batteries, and control systems for electric vehicles have translated into key technologies for humanoid robots.

While the South boasts a mature hardware supply chain, the North excels in software solutions, particularly in large AI models. As of March 2024, Beijing had released over 115 large AI model products, making it the leading city in China. Major products include Baidu's Wenxin, ByteDance's Cloud Lark, JD's Yanxi, and others, covering various application scenarios such as content generation, intelligent Q&A, IT support, and data analysis.

On the financing side, Beijing-based AI startups led the country in 2023, with over 156 funding events, accounting for 29.4% of all AI-related financing in China, surpassing cities like Hangzhou, Guangzhou, and Chengdu.

Whether it is the hardware supply chain advantages of the Yangtze River Delta or Beijing's rich AI talent pool, these are only the foundational conditions for the development of humanoid robots. The real challenge lies in making robots general-purpose, capable of performing tasks in production lines and home services.

Wang Xingxing noted that while many domestic humanoid robot companies have emerged in the past two years by leveraging open-source quadruped robot technology, humanoid robots require more than just large language models for continued development. Although large language models are a part of the robotics equation, they are not the whole solution.

As humanoid robotics continues to evolve, companies are exploring commercial opportunities in specialized scenarios. For example, Yinhe Universal's humanoid robots are set to undergo testing in real pharmacies and supermarkets later this year, with a focus on nighttime staffing and medicine delivery.

Looking ahead, innovation centers across various cities may adopt collaborative development models, with breakthroughs in specific verticals being shared across regions. Despite the long road ahead, the optimism surrounding humanoid robotics in China continues to grow, driven by both government initiatives and entrepreneurial innovation.