China’s humanoid robot captured the world’s attention with a kung fu flip during a televised Lunar New Year celebration, while Chinese mobile phone maker Honor plans to unveil its first humanoid robot at MWC in Spain.
Robotics was marked as a priority under the country’s “Made in China 2025” plan, but the focus was initially on factory automation rather than humanoid robots. Now, rapid advances in multimodal AI are accelerating so-called embodied AI (autonomous machines that operate in the real world), which officials say could help fill labor shortages and drive productivity gains.
In the early stages of humanoid robot development, Chinese companies are outpacing their U.S. rivals in both speed and production, said Selina Xu, China and AI policy lead at Eric Schmidt.
“China has a more robust hardware supply chain, much of it built through the EV sector from sensors to batteries, and the world’s strongest manufacturing base, allowing companies to iterate much faster than their Western competitors,” Xu told TechCrunch.
As a result, Chinese-made robots are not only cheaper, but companies are also able to release new models faster, Xu said, adding that Chinese giant Unitree shipped about 36 times as many units last year as U.S. rivals Figure and Tesla.
According to a Forbes magazine report published last month, only 13,317 humanoid robots were shipped worldwide last year. This is a very small base for an industry that is expected to nearly double annually and reach 2.6 million units by 2035. (Still, this number should be viewed with caution; the report notes that unit sales and numbers representing demo models and pilots remain unclear, underscoring the nascent nature of the industry.)
China’s Agibot and Unitree are the top humanoid robot manufacturers by shipments in 2025, followed by UBTech, Leju Robotics, Engine AI, and Fourier Intelligence, highlighting the Chinese government’s early dominance in the field.
tech crunch event
boston, massachusetts
|
June 9, 2026
The biggest recent shift has been from “demo-driven excitement” to “operations-driven adoption,” Yuli Zhao, chief strategy officer at Galbot, told TechCrunch. Galbot’s G1 humanoid robot appeared alongside robots from Unitree Robotics, Noetix and MagicLab at this year’s Spring Festival Celebration, China’s annual state-run Lunar New Year’s Eve TV show.
“More and more customers are asking, ‘Can robots work reliably in real-world environments and actually take people’s jobs?’ Its practical appeal is strengthened in China as policies and industrial strategies encourage automation upgrades and manufacturing ecosystems are iterating much faster,” Zhao said.
Zhao added that while the pace of progress is “certainly accelerating” due to increased funding for humanoid startups, “the most durable adoption will occur when they can demonstrate reliable and repeatable value in production and service operations, rather than one-off showcases.”
Still, investment helps, and Chinese robot makers are securing it. Last year, Unitree was valued at around $3 billion after its Series C, and it aims to reach up to $7 billion in a future IPO. Meanwhile, Galbot has reportedly raised more than $300 million in new funding, pushing its valuation to $3 billion, making it one of the largest-ever financings in China’s humanoid robot space.
American companies are also moving beyond flashy demonstrations and focusing on real-world implementation. Additionally, they pursue their own offensive goals. For example, the US Emerging Enterprise Foundation plans to produce 50,000 humanoid robots by the end of 2027.
But China is already targeting a combination of affordable mass-market models and high-end applications, and is rapidly expanding humanoids across industrial, consumer, and rehabilitation sectors, according to a December TrendForce report.
The bottleneck of China’s advantage
It is still unclear where Chinese humanoid companies actually stand when it comes to AI systems and integrated software. The industry is primarily betting on visual-verbal-behavioral models and “world models,” but both technologies are still in their infancy. Xu said Nvidia currently leads the space with an end-to-end humanoid software stack, so it’s no surprise that most humanoid startups in China are powered by Nvidia’s Orin chips. However, domestic chipmakers are developing homegrown alternatives, he said.
But humanoid robot makers are still grappling with fundamental problems. The challenge is to enable robot-based models to predict the “next physical state” a robot will face in an unpredictable environment, much like how large-scale language models predict the next word. But unlike LLM, humanoid robot companies can’t simply scavenge the Internet for training data, Xu said. Therefore, although most rely on simulated environments to generate synthetic data, real-world data collection remains essential.
“Humanoids are still far from autonomy due to the lack of data issue. The hardware is now more advanced than the software. The robot body is much more dexterous than it was a few years ago (though there are reliability issues, as we saw with the robot that broke down in the humanoid marathon), but the brain is still in its infancy,” the analyst said.
Safety is also a big hurdle for humanoid robots. One high-profile incident could spark a public backlash, and China is likely considering how to quickly deploy the technology without rushing too quickly. Further regulation is expected as the industry matures.
Given the lack of data, Zhao believes demand for humanoids will increase first in fairly contained workplaces.
“Early momentum is likely to be in industrial manufacturing, warehousing logistics and retail, where tasks are repetitive, time-consuming, and processes are clear, creating real demand and ideal conditions for humanoid robots to deliver value at scale,” he said.
Other APAC players
The development of humanoid robots is not a competition between two countries. From startups to semiconductor giants, Japan’s robotics ecosystem aims to mass produce humanoid robots by 2027. Long a pioneer through projects like Honda’s Asimo, Murata Manufacturing’s Murata Boy, and SoftBank Robotics’ Pepper, Japan focuses on precision and advanced control. One area that is unique to this country is that humanoid robots are increasingly being used to care for the elderly.
James Riney, CEO of Coral Capital, which invests in Japanese technology companies, believes Tokyo will continue to thrive in the humanoid robot industry. “We believe there are three factors driving the adoption of robotics in Japan: a labor shortage and a desire to reduce dependence on mass immigration. Second, there is a widespread cultural view of robots as friends – Doraemon versus the Terminator. Third, Japan already has an advantage in many parts of the robot supply chain.”
Hyundai Motor Co.’s Boston Dynamics division plans to introduce the new Atlas humanoid to its factories by 2028, producing up to 30,000 units a year in the U.S. as part of its AI-driven robotics push.
Still, for China, government policy, industrial strategy, labor shortages, and private capital are all converging to accelerate the country’s push to develop humanoid robots.
“China’s leadership is best understood as an advantage of speed and scale,” Zhao said. “The ecosystem here compresses the entire cycle of R&D, supply chain, manufacturing, integration, and customer deployment into a very tight loop. This means humanoid companies can move from prototype to real-world deployment faster, learn from real-world operations, and iterate at a pace that is difficult to match elsewhere.”
Source link
