From September 16 to 17, the First Aggregated Intelligence Industry Development Conference (2025) was held in Wuhan Optics Valley. The conference gathered representatives from leading enterprises and heads of research institutes across multiple fields, including intelligent robots, flying cars, and the automotive industry supply chain, to jointly discuss the development of the aggregated intelligence industry.

The plenary session on September 17 was divided into the morning Aggregated Intelligence Industry Chain Forum and the afternoon Aggregated Intelligence Innovative Technology Forum. Zhang Yongwei, Chairman of Chebaihui, stated that robots will experience an "electric vehicle moment" within five years. Xiao Qianzhi, Vice President and Chief Public Affairs Officer of AeroFly, believes that future electric aviation may utilize 80% of the core supply chain from the automotive industry.

Early yesterday morning, 42nd Wave arrived in Wuhan Optics Valley to attend this conference full of substantive content. During the event, we also interviewed key figures in the Embodied AI industry, such as Zhao Zhelun, Co-founder of Vita Power, and Zhang Hanwen, General Manager of Ecology for the Robotics Division at Zhixing Technology, collecting substantial insights and first-hand information. Below is the edited summary based on the speakers' content.

The Path to Scalable Robot Applications

1. Zhang Yongwei: Robots Will Experience an 'Electric Vehicle Moment' Within Five Years

Zhang Yongwei, Chairman of Chebaihui, proposed using an aggregated mindset to promote the integrated and synergistic development of three industries: smart vehicles, Embodied robots, and the low-altitude economy.

He stated that these three industries are deeply coupled in three aspects. First, they share a common technological origin; to a certain extent, they are all different terminal products based on artificial intelligence technology. Second, their industrial chains are largely overlapping, with shared segments exceeding personalized ones. The supply chain overlap between smart cars and Embodied AI robots exceeds 60%, with some enterprises exceeding 70%. Finally, there is a trend toward integration in application areas, even leading to the emergence of integrated terminals spanning land and sky. Therefore, the essence of aggregated intelligence lies in shared technology origins, interconnected supply and industrial chains, and converging applications.

He also mentioned that the current high cost of robots stems from an immature supply chain and the lack of a manufacturing system that supports cost reduction through industrialization. By extending the automotive manufacturing system to the fields of robotics and the low-altitude economy, it is possible to achieve continuous reductions in hardware costs for new domains and significant performance improvements, including in perception, computing power, and models.

Finally, regarding development paths, the robotics sector will likely experience its 'electric vehicle moment' within five years, potentially even faster than the automotive industry did initially, characterized by a decline in manufacturing costs and a substantial increase in sales revenue.

2. Zhao Zhelun: Vbot Expected to Ship 30,000-50,000 Units Next Year

Zhao Zhelun, co-founder of Vita Power, discussed how to define consumer-grade Embodied AI products, along with the technical and sales plans for Vbot.

He first introduced the product definition of Vbot: quadruped robots capable of moving within living spaces and communicating via natural language. The combination of these three features enables a robotic product that does not require remote control.

When asked why quadruped robots were chosen as the first product, he explained that in recent years, grippers, quadrupeds, and autonomous mobility have entered the engineering phase, while bipedal locomotion and dexterous hands still have a long way to go. In terms of user scenarios, Vitae Dynamics' definition is not limited to factories or homes but is divided into the natural world and the human-constructed world. Today, robots are transitioning humans from the era of graphical interaction to an era of natural interaction.

He noted that while many ordinary people show strong interest in robots, converting that interest into purchases remains a difficult process. There are three stages from technology to users. The initial stage is purely technical, featuring many demo videos and academic papers; the second stage involves some productization with good reliability, but the utility for users remains unclear; the third stage, true commercialization, requires clearly articulating the utility and aligning price with value. However, no robotics company has truly reached the third stage of effectively matching utility with price and value.

In the dialogue following the speech, 42nd Wave conducted an in-depth exchange with Zhao Zhelun, co-founder of Vitae Dynamics. He mentioned that Vbot's shipment volume is estimated to reach 30,000–50,000 units next year. He also predicted that the consumer-grade quadruped robot market size will be approximately 70,000–80,000 units in 2026, with Vitae Dynamics achieving a 60% market share. For more details, see the conversation with Zhao Zhelun, co-founder of Vitae Dynamics: Vbot production starts next month, expected shipments of 30,000–50,000 units next year.

3. Zhao Mingguo: Robots Must Emphasize Embodied AI in Legs

Zhao Mingguo, a researcher at the Department of Automation, Tsinghua University, and director of the Robot Control Laboratory, focused his speech on embodied AI for robot legs. He argued that most people understand humanoid robots as working in factories, primarily using their hands and intelligence, often overlooking "walking." The primary task for a humanoid robot is "walking," and during this process, significant intelligence resides in the legs that must be emphasized. Therefore, over the past one to two years, he has been helping the industry establish this concept.

He cited the example of robots playing soccer. In 2023 and 2024, robots playing soccer required someone to follow them to resolve issues; if they fell, a person had to take them off the field, making continuous play impossible. However, today's robots can play on the field without human assistance and can autonomously get back up after falling to continue the match.

He then stated that soccer serves as a standard platform. Robots must answer what the corresponding social applications are and where the verified technologies will be applied. If robots can solve embodied AI problems for legs through soccer, and combined with the resolution of embodied AI problems for arms, this integration could effectively address the basic technology requirements for general-purpose humanoid robots.

4. Ji Chao: Physical AI Will Become an Important Component of General AI

Ji Chao, Founder and CEO of Lingdong General and Chief Scientist of iFlytek Robotics, mentioned in his speech that physical AI will become an important component of general AI with a disruptive attitude.

He stated that leading global large model enterprises have densely released results, and domestic large models have shifted from catching up to innovating, with top players becoming prominent. The disembodied intelligent large model era 1.0, transitioning from LLM to VLM, may enter a convergence phase in the coming years.

In this process, embodied AI technology is regarded as the next generation of disruptive technology, and they defined two standards for practical implementation. The first is labor scenarios: improving productivity efficiency under existing paradigms is an important standard for the landing and scaled implementation of disruptive technologies. For example, robots in automotive assembly workshops can achieve a positive return-on-investment cycle within 18 to 24 months, replacing 1 to 2 workers.

Secondly, the reconstruction of the productivity system during the collapse of existing models under clear trends, such as the redesign of next-generation automotive production lines around general-purpose robots, shifting from a "labor-centric" model to a "robot-centric" one.

5. Lv Peng: The current robot market resembles the state of autonomous driving 6-7 years ago

Lv Peng, Vice President at Horizon Robotics and Head of Strategy, Intelligent Driving Product Planning & Marketing, shared his views on robots and intelligent driving.

He believes that the current robot market is very similar to the state of autonomous driving 6-7 years ago: diverse and flourishing, with the industry still in its early stages, though many projects are in the prototype phase. Smart vehicles also went through a process from diversification to industry convergence, eventually leading to product maturity and large-scale deployment by a few key players.

Moreover, many core technologies in the entire robot industry, such as AI algorithms and cloud infrastructure, have a high degree of overlap with those in smart cars. Therefore, in recent years, many Embodied AI startups have been founded by individuals who previously worked in the autonomous driving sector.

6. Zhang Hanwen: Smart cars and Embodied AI have significant overlaps in technology chains and supply chains

Zhang Hanwen, General Manager of the Robotics Business Division Ecosystem at ZhiXing Technology, discussed the reasons for ZhiXing's entry into the robotics industry and its specific practices.

Zhang Hanwen stated that the core reason for ZhiXing's entry into the robotics sector is the high degree of overlap between intelligent vehicles and Embodied AI in their technology and supply chains. They chose "automatic charging robots" as their first implementation scenario because it is highly relevant to automotive applications and offers a clear commercialization path.

To this end, ZhiXing developed an industrial-grade welding robotic arm in-house, focusing on verifying its load capacity, repeat positioning accuracy, and multimodal perception fusion capabilities. Combined with a medium-to-high computing power robot controller co-developed with Horizon Robotics, ZhiXing created a six-axis collaborative robotic arm for automatic charging scenarios. During a public demonstration at the Munich Motor Show, the arm completed a single plug insertion action in 19.5 seconds.

7. Luo Jian: Improving motor power density and torque density is crucial

In his speech, Professor Luo Jian from Shanghai University shared his perspectives on the application of motors in robotics.

Luo Jian stated that in robotics, joint modules can be divided into two categories: upper limbs emphasize responsiveness, while lower joints focus more on output force. However, motors and other actuating gear mechanisms account for a very large proportion of the robot's total weight, ranging between 30% and 40%. Therefore, further R&D on motors to improve power density and torque density is something we must do.

The difference between robot motors and those used in new energy vehicles lies in the fact that robot motors particularly emphasize being small, agile, and having high power density. Although robot motors are small, their power density is higher than that of new energy vehicle motors, making them harder to manufacture.

Finally, he believed that the motor industry chain is shared, differing only in usage scenarios and focus areas. For robot motors, greater innovation will come from composition, specifically by enhancing performance through material applications and compositional improvements, rather than focusing solely on mechanical strength.

8. Su Ying: Goal is to Become the Apple of the Embodied AI Era

Su Ying, Vice President of Accelerated Evolution, presented her thoughts on the development of the humanoid robot industry during her speech.

Su Ying believes that robots playing soccer represents a comprehensive technical pipeline, encompassing embodied intelligence—from perception, decision-making, and movement to group coordination. This serves as a key focus for Accelerated Evolution in building its developer ecosystem. It is also an excellent model for cultivating developers; currently, teams participating in robot soccer competitions are primarily from universities, representing future developers.

She further stated that JiaSu JinHua aims to evolve into a platform company in terms of general robot hardware, operating systems, and development tools. The goal is to enable numerous developers to build upon the JiaSu JinHua platform, with significant optimizations planned for development tools in the future.

9. Wang Jingfan: Robots can enter home scenarios within five years

Wang Jingfan, Head of Business Development for JD Group's Intelligent Robotics Division, shared his views on the future large-scale application of robots during his speech.

He noted that intelligent robots are currently on a rapid development path. Intelligent embodied robots will first achieve commercial applications in industrial, logistics, and retail scenarios. As industry applications have already emerged, technological advancements are expected to enable their use in consumer home environments (C-end) within five years.

Low-altitude flight synergizes with the automotive supply chain, promising significant cost reductions

10. Qu Xiaobo: The cost of flying cars will drop to 1 million yuan in the future

In his speech, Professor Qu Xiaobo of the School of Vehicle and Mobility at Tsinghua University, Chief Scientist of the State Key Laboratory of Automobile Safety, and Academician of the European Academy of Sciences, argued that Aggregated Intelligence can significantly reduce the cost of flying cars.

He stated that Aggregated Intelligence can align part of the automotive supply chain with that of aircraft, drastically lowering costs. With mass production potential, the cost of flying cars could drop to 1 million [currency implied, typically RMB in this context], and further reductions are possible as aggregated supply chain costs decrease through volume scaling.

He also noted that globally there are nearly 300 flying car models targeting passenger transport, which will eventually converge into a few mainstream configurations better integrated with the automotive industry. Based on urban mobility needs, multi-rotor aircraft are expected to account for about 20%, compound-wing designs 80%, and tilt-rotor designs will have virtually no market. By improving the weakest components (the 'shortest plank'), overall performance can be enhanced. Additionally, using virtual-real fusion twin platforms to accelerate testing could potentially shorten the R&D cycle from over five years to two years.

11. Xiao Qianzhi: Future Electric Aviation May Utilize 80% of Core Automotive Supply Chains

Xiao Qianzhi, Vice President and Chief Public Affairs Officer of AeroFlight, elaborated on electric aviation and its relationship with the automotive sector in his speech.

He believes that electric aviation and traditional automobiles share strong commonalities in terms of dimensions. From the perspective of technical theoretical foundations, the two are essentially interconnected, with significant potential for supply chain reuse.

However, from the perspective of industrial practice, there will be a Pareto principle in the future, meaning that electric aviation may use 80% of the core supply chain of the automotive industry, but currently it is unable to achieve even 20%. This is mainly because civil aviation always faces airworthiness safety requirements, which traditional automotive industries cannot reach. Therefore, he hopes that traditional automotive supply chain manufacturers can consider transitioning to electric aviation and jointly overcome airworthiness requirements.

Furthermore, he stated that Aurica Aerospace (WoFei ChangKong) is currently focusing on three aspects of electric aviation: first, aviation-grade safety; second, automobile-grade comfort; and third, consumer-grade cost. These represent three critical challenges facing the electric aviation sector.