At the 2026 China International Fair for Trade in Services, the "Physical AI and Embodied Intelligence Robot Innovation Ecosystem Exchange" event took place in Beijing on September 11th. Chen Jianquan, the Investor Relations lead at UBTECH ROBOTICS, attended the forum and participated in the panel discussion. Below is the transcript of the dialogue.

Moderator崔轲迪: Please introduce yourself, Mr. Chen.

Chen Jianquan: Thank you, Mr. Cui, for moderating, and thanks to the CIFTIS committee for the invitation. I am Chen Jianquan, in charge of investor relations at UBTECH ROBOTICS. The company has been deeply engaged in humanoid robots for more than 14 years, entering its 15th year next year. We successfully listed in Hong Kong in 2023. In the first half of this year, our robotics revenue exceeded 1.2 billion yuan, and we anticipate a strong full-year figure of over 4 billion yuan. We are quite proud to be a pioneer in the industry without becoming a casualty, as many humanoid robot companies that started with us have disappeared along the way. We continue to persist in this field, and we are thrilled to witness the flourishing growth of China's humanoid robot supply chain and applications in recent years. We also appreciate the significant push from primary market investors and secondary market participants who have driven capital and industry momentum. Today, we're here to exchange ideas with industry leaders.

Moderator崔轲迪: We'd like you to share your perspective, which we've heard is quite distinct. By the way, I thought UBTECH ROBOTICS went public in 2024, didn't it?

Chen Jianquan: It was at the end of 2023. Since our listing, the primary market has truly ignited in enthusiasm. As a listed embodied intelligence company, how do we view the differences between the primary and secondary markets?

Chen Jianquan: When we were preparing for the IPO back in 2018, upon my arrival at UBTECH ROBOTICS, and eventually listing in Hong Kong in 2023, the market conditions were not particularly vibrant. However, after listing, we successfully joined the Stock Connect program, attracting mainland capital. As the sole pure-play humanoid robot stock at the time, southbound funds pushed our share price from around HK$90 to a historical peak of over HK$300, boosting our market cap to over HK$130 billion. This clearly highlights the immense market enthusiasm for humanoid robots and the broader tech sector. Looking at the fundamental drivers, we see several key factors: national policy support for the humanoid robot industry, the launch of Tesla's Optimus, and the advancement of large language models like ChatGPT. These have combined to propel the entire Chinese supply chain from both policy, software, and hardware perspectives. Before 2022, if we were building complete robots, every component had to be self-developed and custom-manufactured, with virtually no economies of scale available to upstream suppliers. Our harmonic drives, motors, and cameras were extremely costly, with the total cost of a single robot potentially exceeding $100,000. But thanks to the supply chain development over the past 2-3 years, you can now buy a robot for 30,000 yuan, and even smaller ones for just a few thousand yuan. This progress has driven the growth of entire robot companies and their applications. A remarkable illustration of this is the evolution of the World Robot Conference (WRC). In 2022 and 2023, it was dominated by industrial robots from companies like ABB, KUKA, and Yaskawa, showcasing mechanical arm precision and dexterity tricks like tracing around cans of Coke. This year, however, those are gone, replaced by over 100 humanoid robot companies. In 2023, UBTECH ROBOTICS was the sole one. This rapid expansion over the past few years has spawned enormous growth across the supply chain. What we observe is that as upstream costs decline, terminal costs follow, and subsequently prices for end-user applications drop. We foresee a trend where hardware costs for complete robots will continue to decrease, while advancements in large models will provide even greater intelligence to the embodied intelligence sector. However, we believe this is still in its early stages. Even with foundation models reaching trillion-parameter scales, their understanding of human physical behavior from internet data is just beginning. Translating that comprehension into the physical world will require more time.

Moderator崔轲迪: From your perspective, do you view the newly established embodied intelligence companies in the primary market or those heading for IPOs as competitors for funding?

Chen Jianquan: We firmly believe the pie is expanding. Without these additional enterprises, we couldn't have grown the market ourselves. The influx of companies attracts more capital, which doesn't detract from our market share but rather enlarges it. The supply chain's evolution is a clear testament—if only UBTECH ROBOTICS existed, the industry wouldn't have developed. Even with just UBTECH ROBOTICS and Tesla, it might not have been enough. It's only when downstream orders increase that upstream suppliers are motivated to invest in technology. We sincerely hope that the entire industry—whether companies are preparing for listing or already public—will invest more resources into continuous hardware iteration and collaborate to enhance the applications. We are confident that the embodied intelligence market will eventually be more than ten times larger than the new energy vehicle sector. Particularly in interactions with humans and deployments in commercial and home environments, we agree with Musk's assertion that there will be over 10 billion robots in our midst in the future. It might be hard to imagine now, just as three years ago no one envisioned robots dancing, jumping high, or even breaking Bolt's sprint records. For instance, the Tiangong robot by UBTECH ROBOTICS and other ventures can complete a 100-meter dash in 8.6 seconds, which was inconceivable last year.

Moderator崔轲迪: From a commercialization standpoint, which sector do you prefer?

Chen Jianquan: As an industry player, we prioritize the market space that specific application scenarios offer. Mr. Cui, you noticed that the three previous speakers gave somewhat different views, but fundamentally, they all address the same core: solving human needs.

Moderator崔轲迪: From a market size perspective, there's a concrete scale. However, from initial R&D to actual market penetration at that scale, there's a timeline. It might be a trillion-level market, but even getting to a 100 billion level could take three years when considering the time dimension.

Chen Jianquan: Indeed, going from zero to one is extremely challenging. When we initiated our industrial robotics project in 2023 and only started generating revenue in the latter half of last year, we essentially "worked for free" for major factories from 2023 to the first half of 2025. We invested personnel and machines to collect data and train models, without income. Yet, we remained resolute, confident that solving human needs would inevitably create a market. In industrial settings, with roughly 200 million workers in China, we see significant opportunities in basic workstations. We're starting with tasks that don't require high cycle times, where speed isn't critical, and where traditional robotic arms are unsuitable. Our target clients are primarily top-tier Fortune 500 companies that have already optimized their automation. Through these deployments, we've gradually expanded into scenarios like machine tending and palletizing/depalletizing. The market for even these simple use cases could reach a trillion yuan. Essentially, what we're all discussing is addressing human requirements. In industrial settings, it's about addressing labor needs. In the consumer space, consider DJI's drones—there was no inherent need for drones, but once the technology matured, they fulfilled a different need for photography. As technology advances, we see outcomes like drone-based seeding, heavy lifting, and even life-saving rescue operations during floods in Guangxi. These are all fruits of technological progress. The same trajectory applies to humanoid robots. As models mature and we solve initial needs, we must also consider emotional requirements at home—beyond folding clothes, picking up laundry, and cleaning. Could robots potentially simulate family members to provide daily emotional support? From companionship for children and the elderly to even acting as a stand-in relative with your own personality and traits to comfort your family when you're apart—these are extensions that technology could enable, and we're open to exploring them. That's why we launched our emotional companion robot this year once the technology matured. We also fully concur with Mr. Zou's point that cash flow is absolutely vital; it's a critical foundation for sustainable growth.

Moderator崔轲迪: Is the period between starting R&D and achieving the first commercial revenue particularly agonizing?

Chen Jianquan: Extremely so. Humanoid robots only truly began generating revenue in the second half of 2015, when we delivered the first batch of 1,000 vehicles in accordance with regulations. Before that, our income came from technology spin-offs. We applied our vision and navigation technologies to AGVs, logistics vehicles, and smaller C-end humanoids, including products like self-cleaning cat litter boxes. We had to keep the company alive, build cash flow, and use that to partially fund our R&D. We've found that high-level technology, when applied to lower-tier products, is exceptionally competitive. Take the cat litter box—we started in 2022, and we're now the market leader in Europe.

Moderator崔轲迪: You're not shy about mentioning that, are you? That you have a commercial cat litter box product?

Chen Jianquan: We're not shy about it. It's been crucial in sustaining the company through tough times before humanoid robot technology matured. Even now, it continues to generate steady cash flow. Though its revenue growth isn't as fast as humanoid robots, it has provided tremendous support throughout our journey, and we're very proud of that. [NOTE: The original text includes a promotional line for Sina's futures brokerage platform, which has been omitted per instructions.]