为什么说公共服务业是人形机器人快速实现大规模商业化的应用场景之一？

Currently, major Chinese manufacturers of humanoid robots include Kepler, Hanson Robotics, Fourier, and Ubtech, among others. Notably, Kepler's humanoid robot, particularly its pioneering model K1, represents a significant breakthrough in the field of domestically produced Chinese humanoid robots. The release of this robot marks important progress for China in high-end robotics technology. Fourier Intelligence initially focused on rehabilitation robotics but has leveraged its core technological capabilities to enter the humanoid robot market in recent years. Fourier has launched a series of products including the General Robotic Model GR-1, GR-Pro, and GR-1 Lite.

There remains a gap between domestically produced Chinese humanoid robots and Tesla’s Optimus. A notable advantage of Tesla’s Optimus is its integration with the Full Self-Driving (FSD) system. However, the gap in intelligent technologies between China and the US is gradually narrowing. Meanwhile, China’s humanoid robot industry demonstrates clear advantages in supply chain integration, mass production capabilities, and application scenario implementation.

I think the cost of humanoid robots is expected to drop to below 200,000 RMB within the next five years, and by 2030, it is anticipated to fall further to below 100,000 RMB. In the core components of humanoid robots, the cost of precision reducers accounts for approximately 36%, which is the highest proportion. Currently, the global market for precision reducers is mainly dominated by Japanese companies, but in recent years, the market share of domestic precision reducers in China has been steadily increasing, which will significantly reduce the future costs of humanoid robots.

I believe that it is likely to be at least 20 to 30 years before humanoid robots can achieve large-scale commercial application. One key step towards this goal is for humanoid robot companies to reach true mass production scale, which for China may take about 10 years to accomplish. Currently, many countries and regions around the world, including China, maintain a cautious attitude towards the large-scale commercial application of humanoid robots. While technology continues to advance, there are still numerous challenges to overcome, such as reducing costs, improving technical maturity, and enhancing social acceptance.

Personally, I believe that industrial and public service scenarios will see the fastest large-scale commercialization of humanoid robots.

In industrial settings, first, humanoid robots can replace human workers in performing dangerous or highly repetitive tasks, leading to long-term cost savings and increased productivity for businesses; second, industrial environments typically offer a more closed and controlled work setting, which facilitates safe operation and efficient deployment of humanoid robots; third, the demand for humanoid robots in industrial contexts is very specific, such as material handling, assembly, and inspection, which allows for targeted optimization of the design and functionality of humanoid robots. Manufacturers like Tesla are introducing humanoid robots into factory environments to perform tasks such as torque tightening, flexible assembly, and material handling.

In public service settings, first, the demand is widespread, such as in restaurants, hotels, elderly care facilities, and hospitals, where substantial human labor is required to provide services and support, and humanoid robots can substitute or assist human employees in performing repetitive and labor-intensive tasks; second, the level of standardization is higher compared to the highly customized needs in industrial environments. Public service scenarios often involve more standardized tasks, meaning that costs can be reduced, and efficiency improved through batch production and programming; third, the level of technological maturity, current humanoid robot technology can already execute specific tasks in relatively simple environments. Public service scenarios typically do not require overly complex operations, providing an opportunity for early commercialization of humanoid robots. For example, during Expo 2020 Dubai, Ubtech's humanoid robot Walker provided services for six months at the China Pavilion, conducting over 1,000 receptions and tours, totaling more than 1,200 hours of service. This was the first successful case of commercial application of a humanoid robot. In 2022, the Walker X was exported to Saudi Arabia to provide intelligent services locally.

I personally disagree. While there have been significant advancements in the hardware technology of humanoid robots, there are still considerable gaps in legal frameworks and ethical considerations. For instance, when a humanoid robot causes harm, how should responsibility be allocated? Should it be the manufacturer, the owner, or the robot itself? Establishing an effective mechanism for accountability is one of the crucial prerequisites for the large-scale commercialization of humanoid robots. Secondly, humanoid robots may collect vast amounts of personal information, so measures need to be taken to protect users' privacy rights. Using humanoid robots for surveillance activities could be seen as an infringement of personal privacy. Furthermore, the proliferation of humanoid robots might have a significant impact on the job market, potentially leading to the disappearance of certain types of employment. When Apollo Go, an autonomous driving vehicle service, was launched, some ride-hailing drivers expressed similar concerns.

According to our institute's projections, the market size for humanoid robots in China is expected to reach 452.4 billion yuan by 2030, with an average annual compound growth rate of 84%. Ten years from now, humanoid robots are expected to enter mass production, with a market size surpassing 800 billion yuan. As the technology matures and costs decrease, humanoid robots will demonstrate significantly improved environmental adaptability and human-machine interaction capabilities in complex and open consumer scenarios (C-end). Demand in the home environment will increase, with applications primarily focusing on companionship and service, characterized by high levels of personalization and customization. Additionally, against the backdrop of an aging population, declining birth rates, and rising labor costs, humanoid robots exhibit significant potential across multiple sectors.

Current policies focus on establishing a comprehensive innovation system, ensuring breakthroughs in key technologies such as "brain, cerebellum, limbs," and bringing whole machine products to international advanced levels while achieving mass production. On this foundation, a healthy enterprise and industrial ecosystem is being built, along with a stable supply chain. Finally, through demonstration applications of humanoid robots in special, manufacturing, and civilian service scenarios, effective governance mechanisms and methods are being explored to continuously expand their application areas.

From the perspective of existing policies, additional encouraging measures can be implemented to promote the development of the humanoid robot industry. First, financial and investment support aims to further increase investment and financial support for the humanoid robot industry, especially by providing long-term loans and venture capital to accelerate technology research and development and industrialization. Second, talent introduction and cultivation aim to establish more talent training programs and scholarship plans to attract top global talent to join China's humanoid robot research and industrial construction. Additionally, regulatory policies aim to reduce improper operations and ethical issues to promote the healthy development of the industry.

China holds a leading position in six-axis force sensors and is rapidly advancing in dexterous hands, while closing the gap with international leaders in actuators and ball screws. Taking sensors as an example, the global six-axis force sensor market is primarily composed of companies from Europe, America, Japan, South Korea, and China, with the number of participating Chinese enterprises steadily increasing in recent years. Regarding ball screws, in 2022, mainland Chinese companies held a 30% share in the mid-range ball screw market and a 5% share in the high-end market. In terms of actuators (motors), domestic companies such as Kinco has introduced FMC series motor products that align with international standards, achieving product size standardization and offering a clear cost advantage. Another company, Guangzhou Haozhi Industrial Co., LTD., focuses on the development of frameless torque motors, whose products can achieve torque fluctuations of less than or equal to 1% and an overload capacity of 3.5 times, providing strong power support for the high-performance operation of humanoid robots.

In the realm of end-to-end AI technology, I believe China is progressing alongside international counterparts. Tesla stands as a representative enterprise of end-to-end technology, particularly in its Full Self-Driving (FSD) suite. In China, end-to-end autonomous driving has become a clear industry consensus. According to interviews with front-line experts in the autonomous driving industry, 90% stated that their respective companies have invested in the research and development of end-to-end technology. Meanwhile, some leading companies within the industry are indeed pushing forward the commercial application of end-to-end autonomous driving technology. It is expected that modular end-to-end systems will begin to be deployed in actual vehicles by 2025. At this year’s Chongqing Auto Show in China, the CEO of Li Auto proposed a technical framework known as "end-to-end" combined with VLM (Visual Language Model) and generative verification systems. This suggests that we are not only closer to realizing the dream of true autonomous driving but also likely to see it become a key architecture for future robotics technologies.

关于头豹