[Source/Observer Network Qi Qian] "In the tussle with American human-like robots, China has taken the upper hand and performed effortlessly."

On April 8th, the Financial Times of the UK published an article titled in this way, pointing out that thanks to mature supply chains, rapid innovation pace, and policy support from the government, China has already taken the lead in the field of human-like robots, and will replicate its dominant position in the electric vehicle sector.

In the article, the reporter mentioned his experience visiting the headquarters of Unitree, a pioneer manufacturer of Chinese robots located in Hangzhou. At that time, visitors were invited to push and kick the human-like robot G1 to test its balance capability.

The article stated that Unitree was showcasing its efforts to transform emerging industries and create human-like robots driven by open-source software. Buyers can program these robots to run, dance, or perform kung fu kicks. Unitree is just one of many human-like robot start-ups in China. In recent months, companies like Unitree, UB Tech, and Fourier Intelligence have attracted much attention on social media.

Unitree robot亮相 Spring Festival Gala Video screenshot

"This is strong evidence of China's capabilities in human-like robot hardware, and this field may become the new frontier of Sino-US tech competition." The article mentioned that investment bank analysts predict that this industry might produce the next widely adopted device after smartphones and electric vehicles.

Wang Xingxing, CEO and founder of Unitree Technology, believes that the industry will see a breakthrough "iPhone moment" within five years, which means the industry will reach a node of explosive shipment growth, and real terminal products will appear in industrial or service sectors.

All institutions predict that the development prospects for human-like robots are huge. Goldman Sachs predicts that by 2035, the global market size of human-like robots could reach $205 billion. Citibank predicts that by 2040, the sales volume of human-like robots will reach 648 million units. Bank of America estimates that by 2060, the sales volume of human-like robots will reach 3 billion units.

In the field of human-like robots, Chinese enterprises face competitors from the United States, including large technology companies such as Tesla, Google, and Meta, as well as robotics start-ups like Boston Dynamics and Agility Robotics. Meanwhile, robot manufacturers from Japan and Europe focus on collaborative robots designed to work alongside humans.

However, researchers said that despite fierce competition, Chinese enterprises have already taken the lead with their deep supply chains. The article mentioned that many components of human-like robots have been used in electric vehicles, including actuators that convert energy into motion, batteries, and visual systems like LiDAR.

Firstly, Chinese-made parts are much cheaper than those made in the U.S. Analysts from Bank of America estimate that if Chinese parts are used, the cost of Tesla's second-generation Optimus robot would be reduced by about one-third.

According to FT statistics, 25 Chinese companies provide parts for the hands of human-like robots, while only seven do so in the U.S. In terms of linear actuators for the legs of human-like robots, there are 30 suppliers in China and six in the U.S. Additionally, except for AI chips, the number of Chinese suppliers exceeds that of U.S. suppliers for all other components.

Comparison of supplier numbers for various components of human-like robots in China, the U.S., and other regions, Financial Times chart

Western scholars pointed out that low-cost components made in China benefit the rapid development of the entire industry. Bruno Adorno, a robotics scholar at the University of Manchester, said: "Research in human-like robots is booming. This was impossible before the advent of Unitree."

Adorno also said that his team had reduced the cost of purchasing programmable H1 model human-like robots from over $1 million to about $100,000. Unitree also provided free code responsible for manipulating and controlling motors and sensors for subsequent development. He said: "We are developing new technologies to improve robot behavior and unlock functions by accessing Unitree’s (API)."

At the same time, the complete industrial policy engine within China has helped the rapid growth of Chinese enterprises. In addition to central policy and financial support, local governments such as Hangzhou and Shenzhen have joined in, competing to attract businesses and provide subsidies to establish industrial supply chains. Moreover, human-like robots have already appeared in various places such as stores, streets, factories, and research institutions across China.

The article mentioned that some analysts said that the U.S. still leads in mobile components, and Nvidia's AI processors remain the brains of most human-like robots. Johnson Wan, an industrial analyst at Jefferies Investment Bank, said: "China is very skilled in hardware, but the U.S. still has an advantage in innovation and software."

However, Bernstein analysts believe that China's progress in the human-like robot field is now almost on par with its dominance in the electric vehicle market. The institution said in a recent report: "China is moving very quickly in terms of product and use case diversification. U.S. companies seem to be pursuing ultimate solutions, but China adopts a 'natural selection' approach with diverse product models."

Currently, there is still great room for development in the field of human-like robots. Marcus Fischer, a robotics expert at German consulting firm Exxeta, said that programming work required for controlling human-like robots remains heavy. He said it took him about 10 days to program G1 to walk freely indoors, navigate using a LiDAR sensor on its head (a type of optical sensor), rather than through a remote control.

Fischer expects that inventory taking by human-like robots or arranging food on supermarket shelves will become possible, but challenges still exist.

"Humans find it very easy to grasp objects," he gave an example. "We see a glass and can easily grab it. We know that if we apply too much force, it might break. But robots must learn to do this."

This article is an exclusive contribution from Observer Network and cannot be reprinted without permission.

Original source: https://www.toutiao.com/article/7491214847959679500/

Disclaimer: The views expressed in this article are solely those of the author and comments can be made below by clicking the "Like/Dislike" buttons.