How do investors conduct a post - event review of the Robot Sports Games?
In the recently concluded world's first humanoid robot sports games, 280 teams from 16 countries across five continents competed in 487 matches in a total of 26 events, including competitive races, exhibition matches, scenario-based competitions, and peripheral races, resulting in 26 gold medals.
From the intense competition on the track and field to the interesting confrontations in football, combat, and other events, and then to the scenario-based competitions that are close to real-world applications, leading humanoid robot companies such as Beijing Humanoid Robot Innovation Center, Galaxy Universal, Unitree Technology, Acceleration Evolution, Songyan Power, Xingdong Era, Xinghai Map, and Qianjue Technology demonstrated the technological breakthroughs and development potential of humanoid robots in areas such as full autonomous navigation, motion control, and scenario adaptation.
However, the significance of this sports games goes beyond the exciting moments on the field. To review the event, a reporter from Venture Capital Daily interviewed numerous robot companies, universities, research teams, and investors, attempting to present the most cutting-edge industrial observations.
A review after the games reveals several obvious trends: At the technical level, humanoid robots still have deficiencies in dynamic balance, environmental perception, and multi-robot collaboration. Frequent incidents such as falling, movement errors, and slow reactions occurred during the competition, but this also clarifies the direction for subsequent research and development.
Many participating teams told the reporter from Venture Capital Daily during post-game exchanges that they will increase investment in sensors, algorithms, and hardware design in the future. At the same time, full autonomous navigation and motion control are generally regarded as the key points of research and development. The performance of Jushen Tianong Ultra, which completed the entire 1500-meter race independently without human remote control, has shown more teams the feasibility and potential of this direction.
From a commercial perspective, the robot sports games have also shown their value: The event attracted traffic and sponsorship, strengthened the linkage with the education market, and the "sports games economics" is gradually taking shape. The games also intuitively demonstrated the potential of humanoid robots in industries such as manufacturing, medicine, hospitality, and warehousing, encouraging more companies to invest and accelerating their entry into the market.
Full Autonomous Capability: From "Remote Control" to "Throwing Away the Remote Control"
The track and field running competition was undoubtedly one of the most popular attractions of this sports games, and the teams from Unitree and Jushen Tianong Ultra almost dominated the top three positions in this category of events.
Among them, "Jushen Tianong Ultra" was the only participant in the 1500-meter race who completed the entire race independently without human remote control.
Wei Jiaxing, a relevant person in charge of the Beijing Humanoid Robot Innovation Center, told the reporter from Venture Capital Daily that this is another upgrade for Jushen Tianong Ultra after using UWB semi-autonomous navigation in the robot marathon. It has achieved real-time environmental understanding and autonomous motion planning in complex dynamic environments.
"On the path of the evolution of embodied intelligence, throwing away the remote control is not just a simple choice of functions but a leap in whether a robot can truly have an autonomous 'brain'. The essence of remote control operation is to let the robot temporarily borrow the human brain, where the operator replaces the robot in environmental perception, path planning, and assists in dynamic balance control. The full autonomous mode means that the robot needs to complete the perception-decision-execution closed-loop on its own and autonomously handle possible emergencies." Wei Jiaxing analyzed.
Regarding the performance of Tianong in this sports games, Guo Yijie, the person in charge of motion control at the Beijing Humanoid Robot Innovation Center, told the reporter from Venture Capital Daily that the results met the expectations. He mentioned that different plans were prepared for different scenarios in advance, and an autonomous mobile plan was adopted to track the runway line in the track scenario.
Regarding the key and difficult points in the industry's research and development of full autonomous navigation for humanoid robots, Guo Yijie talked about the breakthrough directions and key progress of the team in technical research.
He pointed out that the team's primary task is to tackle two core technical problems: First, to achieve stable recognition of the runway line in a high-speed moving state and solve the problem of the accuracy of visual perception in dynamic scenarios; second, to achieve stable tracking based on recognition, which requires the robot to have the coordinated ability of high-speed movement and stable direction control to ensure the accuracy of posture and trajectory during the movement process.
In addition to the above core problems, the team also achieved autonomous movement of the robot in complex terrains through the integration of perception and movement algorithms. At the same time, by combining real-time physical perception and semantic perception technologies, it achieved mapless semantic navigation, got rid of the dependence on preset maps, and improved the adaptability of navigation scenarios.
Guo Yijie especially mentioned that the accumulation at the algorithm level is an important advantage of the team. Relying on the early start of research on autonomous navigation algorithms for humanoid robots, the team developed a more adaptable algorithm system in line with the unique motion structure and gait characteristics of humanoid robots, which provided key support for technical breakthroughs.
When talking about future development, he said that the team will continue to promote algorithm iteration to further enhance the stability and reliability of the robot in complex terrain movement and mapless navigation scenarios and continuously promote the practical application process of full autonomous navigation technology for humanoid robots.
The Langyi robot, which participated in the track and field competition of this sports games, also used the full autonomous mode to participate. It told the reporter from Venture Capital Daily that Langyi relies on its self-developed embodied cerebellum module to achieve full autonomy without remote control. This system can include each joint of the robot's whole body in the obstacle avoidance control scope, achieving autonomous walking and stable obstacle avoidance without human remote control.
The industry generally believes that full autonomy without remote control is an important development trend in the field of humanoid robots. Because the ultimate goal of humanoid robots is to become "general laborers", and only by achieving full autonomy without remote control can they flexibly perform diverse tasks in various complex real scenarios and fully realize the value of general laborers.
The Langyi robot team judged that full autonomy without remote control is a major industry trend, but the technical difficulty and threshold of autonomous navigation for humanoid robots are high, posing high requirements for the R & D investment of enterprises. Humanoid robots need to be applied in various industries, and the real scenarios are highly diverse and complex, making the navigation of humanoid robots much more difficult than autonomous driving.
Wang Xingxing, the founder of Unitree Technology, also said at the scene of this sports games that Unitree's H1 actually has the ability to operate without a remote control, but due to different strategies, in order to pursue speed, Unitree still adopted the remote control strategy in the competition. "We will definitely use the autonomous mode in the next competition."
Xu Biao, an investor in the robot field at Shoucheng Holdings, told the reporter from Venture Capital Daily that "full autonomy without remote control" is the key dividing line for the real implementation of embodied intelligence technology. Although in the current competition environment, remotely controlled robots often have an advantage in speed performance, the full autonomous system represents the complete closed-loop ability of the robot's perception-decision-execution, which is the necessary prerequisite for achieving real commercialization and large-scale application.
A coefficient of 0.8 was set in the competition to "compensate" the results of the autonomous system, which is to some extent a recognition of the technical difficulty and industrial value. After all, in real scenarios, no one can remotely control a robot in real-time to greet guests, deliver goods, or perform operations. "Only robots that can complete tasks autonomously can truly integrate into urban services, manufacturing factories, and family scenarios and become part of the new productive forces." Xu Biao said.
Demonstration of Robot Motion Control Abilities in Various Fields
The reporter from Venture Capital Daily noticed that in addition to track and field, robots from different companies demonstrated their respective motion control advantages in multiple events.
Unitree Technology stood out with its high-speed and stable running ability; Acceleration Evolution took charge of the entire football event, showing strong confrontation and collaborative control abilities; Songyan Power demonstrated its flexible control and dynamic imitation abilities in events such as artistic gymnastics, long jump, and dance; the L7 of Xingdong Era broke the record in the high jump event, and its reinforcement learning algorithm realized the dynamic adjustment of joint torque to complete difficult actions such as "zero-radius turns".
Xu Biao analyzed that overall, each company had its own highlights and strengths. This was a very successful technology demonstration and industrial linkage, and almost every event showed the improvement and breakthrough of domestic robots in motion control.
When talking about the current level of Chinese humanoid robots in motion control, he believes that Chinese humanoid robots have global competitiveness in the field of motion control, especially in aspects such as gait generation, balance control, and action reproduction. However, there is still room for improvement in the generalization ability of the control system, the efficiency of energy consumption control, and the migration stability in complex heterogeneous scenarios.
So, as an investor, how much weight do you think the motion control ability should account for in the valuation model of humanoid robots?
Xu Biao believes that it depends on the company's positioning. For platform-based humanoid robot companies, "we will set the weight of motion control ability at about 30% - 40% because it directly determines the upper limit of universality and platform ability; for companies targeting specific scenario applications, the weight will be between 15% - 20%, and we will pay more attention to their 'operation + perception' ability bound to tasks."
The Langyi robot team analyzed to the reporter from Venture Capital Daily that from the perspective of the sports games, some companies made in-depth customization and simulation of hardware and motion control algorithms to improve the performance of their robots in the competition, enabling the robots to perform well in events such as running, gymnastics, and long jump.
"This event is of far-reaching significance for motion control. It not only comprehensively demonstrated and tested the technical levels of each company but also allowed companies to see the gaps and make improvements. The diverse events promoted innovation in multiple disciplines such as mechanical structure and artificial intelligence, driving the iterative upgrade of motion control. At the same time, the event attracted more companies and research forces to participate, accelerating the overall implementation of the industry."
Sports Games Economics: From "Performance" to "Business"
In the performance session, Acceleration Evolution and Unitree became the focus of the whole event. In the 5V5 football game, the AI players passed the ball accurately; in the combat ring, the two robots fought with sharp moves, and there was even "provocation" after a fall, creating a lively atmosphere on the scene. Behind these scenes, it reflects the huge market space that humanoid robots are opening up.
Under the entertainment shell of the robot sports games, there is considerable commercial value hidden. Cheng Hao, the founder of Acceleration Evolution, revealed that the robot football league hosted by the company in June this year was fully booked, with a network traffic of nearly 300 million, and it has attracted sponsors to participate.
The High-Tech Robot Industry Research Institute predicts that by 2030, the market size of Chinese humanoid robots will reach 38 billion yuan, and the sales volume will soar from 4,000 units to 271,200 units, with an annual compound growth rate of over 60%, resonating with the popularity of the event.
Amid the doubts of "no clear implementation scenarios", the sports games are opening up a unique "sports games track". Events such as football, dance, and combat not only demonstrated technological breakthroughs but also made the "sports games" itself a scenario where robots can be applied.
Zhao Weichen, the vice president of Acceleration Evolution, said that the sports games are both a "testing ground" for technology and an amplifier of commercial value. The requirements of the event force the iteration of core technologies, while lowering the threshold for public understanding, attracting sponsors, and giving rise to new forms such as event IPs and exhibitions.
Zhao Weichen also mentioned that most of the participating teams in this football game were from top global universities. They developed algorithms and agents based on Acceleration Evolution's hardware platform and put forward requirements for the system openness. This not only promoted the software iteration but also made it possible for the motion, perception, navigation, and decision-making abilities trained in the football scenario to be transferred to real scenarios such as family companionship and industrial operations.
Some investors bluntly said that robot football has become a complete commercial track: tickets, copyrights, sponsorship, player transfers, and youth training are all available. "There is no need to look for scenarios because it is a scenario in itself."
The same is true for the combat event. When two 45 - kilogram robots fight with each other at a speed of 12 m/s in the octagon cage, what the audience consumes is no longer a "mechanical performance" but competitive content similar to human combat - with weight classes, rankings, and commentary. Tickets and live - broadcast copyrights can be directly monetized. At the same time, combat puts forward extreme requirements for instantaneous torque, impact resistance, and fault self - repair, and it also pushes cutting - edge achievements such as motors, servos, materials, and batteries to the verge of mass production at once. In the future, it can be applied to industrial arms, logistics vehicles, and even electric vehicles, similar to the application of F1 racing technology.
Scenario Implementation: "For Real" in Factories, Hotels, and Hospitals
In the scenario - based competition of this robot sports games, four scenarios were carefully set up: factories, warehousing centers, hospitals, and hotels. The selection of these scenarios was not accidental. They are currently the fields where humanoid robots are most widely applied and the actual demand is the most urgent. In factories, robots can assist in completing repetitive tasks such as precision assembly and material handling; in hotels, they can provide personalized services such as room service and guest guidance.
The setting of the scenario - based competition not only provided a competitive stage for robots close to real - world applications but also intuitively demonstrated the practical value and development potential of humanoid robots in various fields. Most of the teams that won medals were robot companies that had achieved commercial implementation in real scenarios, and their daily technical accumulation laid a solid foundation for their final victory.
In the final of the "Scenario Competition - Hospital Drug Sorting Skills Competition", the Galaxy Universal Galbot team won the championship with a time of 10 minutes and 22 seconds. Galaxy Universal told the reporter from Venture Capital Daily that this event was highly competitive, and the difficulty of autonomous operation was extremely high. The competition focused on the application of humanoid robots in the medical field, requiring the robot to autonomously recognize prescriptions and complete sorting. Throughout the process, the Galbot team achieved full - autonomous operation without any remote control. Galaxy Universal said that sorting is just the starting point, and in the future, it will continue to promote the implementation of autonomous robots in