Ni Tao is IE’s bi-weekly columnist, giving exclusive insight into China’s technology and engineering ecosystem. His Inside China column explores the issues that shape discussions and understanding about Chinese innovation, providing fresh perspectives not found elsewhere. To get to read every edition of Inside China and our other exclusive content, subscribe to IE+.
Two recent news items have provided a shot in the arm to China’s burgeoning humanoid robotics industry.
First, on May 31, UBTECH, a pioneer in this sector, signed an agreement with Dongfeng Motor, a major state-run automaker, to supply the plant in Liuzhou, southwestern China, with its iconic Walker S humanoids.
These robots will replace human employees by conducting inspections, filling fluids, and assembling auto parts.
However, the exact number of Walker S units to be deployed and the specific facilities remain undisclosed, as Shenzhen-based UBTECH declined to provide details to the media.
In a separate development, the robotic arm of China’s tech titan Xiaomi Group announced its relocation to the Xiaomi smart car factory in a sprawling robotics-themed industrial park in suburban Beijing.
According to a statement from Xiaomi’s robotic affiliate, the corporation plans to roll out its bionic humanoids for use in its own production bases.
In the foreseeable future, Xiaomi’s robotic products, including its full-sized humanoid CyberOne and quadruped CyberDog, will transition from specialized robots for manufacturing-specific purposes to general technologies for broader applications.
Representative image.
On the ground reality
These developments look promising until you delve deeper for a reality check.
Those following Chinese humanoid robotics might recall that barely three months ago, UBTECH’s iconic Walker S robot entered a factory of NIO, a Chinese EV maker, for a series of “onsite training” exercises.
The training involved the robot autonomously walking the length of production lines, planning its path, and avoiding obstacles, among other tasks.
However, Chinese media reports cited insiders in February, stating that this training was merely part of a pilot to explore applications, not a real delivery or mass rollout.
Perhaps UBTECH’s robot gained enough experience during the NIO factory warm-up to prepare for real-life assembly line jobs? The answer is probably no.
The latest partnership with Dongfeng is likely a repetition of previous practices or a pilot on a larger scale. Either way, it can hardly be taken seriously.
Chinese press stories partly support this conjecture. Insiders suggested that the local government brokered the tie-up, making it more symbolic than substantial. Additionally, the two sides have just begun their cooperation, with no plans for the future.
Zhou Jian, founder, and president of UBTECH, boasted to reporters at the time of its listing on the Hong Kong Stock Exchange late last year — after racking up a cumulative loss of 2.4 billion yuan ($331 million) over three years — that the firm was in talks with several EV makers and aimed to collaborate with more manufacturing clients in 2024.
Nonetheless, the prospects of the loss-making firm are less rosy than he depicts. UBTECH’s 2023 full-year financial report indicates that its education and intelligent robot businesses recorded 347 million yuan ($47.89 million) in revenue, down 32.8 percent yearly.
Although increased sales of consumer-grade and logistics robotics compensated for the dwindling income in 2023, the category for which UBTECH became famous—smart robotics, including its Walker series—continues to struggle financially.
Embarrassingly, the company has sold only 10 Walker robots in three years, contributing a mere 16 million yuan ($2.21 million) to its overall earnings.
One cannot help but wonder how long UBTECH can continue to burn cash and reverse losses for its humanoid robotics.
The plight of UBTECH is telling, for it mirrors the contrast between China’s rising fortunes in the humanoid robotic space and its general lack of clear monetization options.
Are investors betting on a mirage?
Over the past two years, many up-and-coming players have ridden the wave of humanoid robotics.
For example, Agibot, a Shanghai-based humanoid robot developer, has been showered with venture funding and favorable policies. Local state-run conglomerates pledged to open their factories to Agibot for the onsite deployment of its bipeds.
High-value-added segments like automobiles, electronics, and semiconductors could be major application scenarios for humanoid startups like UBTECH.
However, these scenarios are likely to remain the same in the next few years.
The reasons are myriad. In automaking, humanoids are limited to relatively low-end jobs like unmanned inspection, car frame checks, quality screening, and small parts assembling.
The truly valuable parts of carmaking, such as pressing, welding, and paint-spraying, are mostly assigned to industrial or collaborative robots. Humanoids are not cut out for this type of work.
In sectors like 3C, photovoltaic, lithium battery, semiconductor, and 3PL, the big patrons of robotics in China, autonomous mobile robots (AMRs) are central to the entrenched automation solutions.
In warehouses and factories, AMR fleets often shuttle between workstations, handling tasks such as order picking, material handling, and transferring.
Humanoids don’t stand a chance in competing with AMRs. After all, who needs a 1-million-yuan humanoid to carry boxes when a 50,000-yuan AMR can do the job better?
In conclusion, humanoid robots currently play a lesser role in manufacturing than specialized robots. To become versatile helpers, they must compete with more cost-effective robotic variants.
Humanoids’ commercial prospects are even bleaker in roles involving close contact with humans. Immature technologies make the talk of using robots as caregivers, butlers, or companions sound like distant fantasies.
Safety concerns also loom large in a home setting, making it hard for humanoids to penetrate the consumer market.
Another barrier is the high price of humanoids, often between 500,000 and 1 million yuan. Chinese firms are highly skilled at supply chain management and cost control, but this is concerned with volume assumption.
For humanoid developers, leveraging economies of scale requires many users to buy their products. This will require more than homegrown innovation alone.
Despite fancy demo videos, many Chinese humanoid startups still need to produce their technologies at a competitive price. This is challenging even for the most deft cost-killers.
Chinese robotic upstart Unitree dropped a bombshell when it unveiled a highly dexterous humanoid called G1 in May this year, with the no-frills version coming in at a jaw-dropping price of less than 100,000 yuan.
Fanfare aside, G1 came across more as a marketing ploy than a realistic production model. So far, no one has purchased the entry-level G1 at that price. Otherwise, reviews would have cropped up all over the internet.
Therefore, it is worth questioning the true financial well-being of these thriving Chinese robotics companies. Despite their bold predictions, can they continue to secure investor support to stay in the business?
Most are private companies, so no financial figures are available for analysis. However, a closer look will reveal that their primary customers will likely be the STEM/education market in the next two to three years.
Given my observations, university robotics departments already use humanoid or quadruped robots for development, application expansion, and competition, and they have the budget to afford these expensive gadgets.
Smoke and mirrors: China’s humanoid robot sector needs to do better
Going forward, this market segment will likely be a main battleground for domestic humanoid robot companies. However, the old scourge of oversaturation is almost inevitable.
With this in mind, much of the glowing reporting on the ascent of the Chinese humanoid robot industry is hype.
Technologically and commercially, Chinese humanoid roboticists have numerous hard nuts to crack – all harder than getting a robot lying prostrate to push itself up into a standing position.
Challenges include an AI talent shortage and a problematic reliance on NVIDIA’s GPUs and Isaac Lab simulation software, risking a crippling trade embargo.
The next two years will be pivotal for the industry. If significant progress is not made and humanoid robots cannot replace valuable human tasks, it will face a shakeup. The capital frenzy will subside, and news like UBTECH robots in car factories will remain feel-good publicity, unable to address immediate needs.
RECOMMENDED ARTICLES
Stay up-to-date on engineering, tech, space, and science news with The Blueprint.
By clicking sign up, you confirm that you accept this site's Terms of Use and Privacy Policy
ABOUT THE EDITOR
Ni Tao Ni Tao worked with state-owned Chinese media for over a decade before he decided to quit and venture down the rabbit hole of mass communication and part-time teaching. Toward the end of his stint as a journalist, he developed a keen interest in China's booming tech ecosystem. Since then, he has been an avid follower of news from sectors like robotics, AI, autonomous driving, intelligent hardware, and eVTOL. When he's not writing, you can expect him to be on his beloved Yanagisawa saxophones, trying to play some jazz riffs, often in vain and occasionally against the protests of an angry neighbor.
RELATED ARTICLES
0