Why China’s humanoid robot industry wins the first market | TechCrunch

Chinese humanoid robots seized global attention with kung fu jumps at the national Spring Festival televised gala, as Chinese phone maker Honor prepares to unveils its first humanoid robot at the MWC in Spain.

Robotics has been designated as a priority under the national agenda “Made in China 2025” planalthough originally focused on factory automation, rather than humanoids. Today, rapid advances in multimodal AI are accelerating what’s known as embodied AI — autonomous machines operating in the real world — a push that officials say could help offset labor shortages and drive productivity gains.

At this early stage in the development of humanoid robots, Chinese companies are surpassing their U.S. rivals in speed and volume, said Selina Xu, head of China and AI policy at Eric Schmidt’s office.

“China has a more robust hardware supply chain – much of it built in the electric vehicle sector, from sensors to batteries – and the strongest manufacturing base in the world, allowing companies to iterate much more quickly than their Western competitors,” Xu told TechCrunch.

As a result, not only are Chinese robots cheaper, but companies can also launch new models more quickly, Xu noted, adding that major Chinese player Unitree shipped about 36 times more units last year than its U.S. rivals Figure and Tesla.

Global shipments of humanoid robots totaled just 13,317 units last year, according to Forbes report published last month. That’s a tiny base for an industry that’s expected to nearly double each year and reach 2.6 million units by 2035. (Still, the numbers should be viewed with caution. The report notes that it remains unclear how many units represent commercial sales versus demonstration models or pilot deployments, highlighting the early-stage nature of the industry.)

The top makers of humanoid robots by 2025 were led by China’s Agibot and Unitree, followed by UBTech, Leju Robotics, Engine AI and Fourier Intelligence, underscoring Beijing’s early dominance in the sector.

The biggest shift recently has been from “excitement about demos” to “operations-driven adoption,” Yuli Zhao, chief strategy officer at Galbot, told TechCrunch. Galbot’s humanoid robot, the G1, appeared at this year’s Spring Festival Gala, China’s annual Lunar New Year television broadcast, alongside robots from Unitree Robotics, Noetix and MagicLab.

“More and more customers are wondering: Can the robot work stably in real-world environments and actually alleviate people’s work? This practical appeal is enhanced in China because industrial policy and strategy encourages automation upgrades and the manufacturing ecosystem makes iterations extremely fast,” Zhao said.

Although increased funding for humanoid startups “has definitely accelerated” the pace of progress, “the most sustainable adoption comes when you can demonstrate reliable, repeatable value in production or service operations, not just a one-time showcase,” Zhao added.

However, investing helps and Chinese robotics manufacturers are succeeding. Last year, Unitree was valued at around $3 billion after closing its Series C round, with ambitions to reach up to $7 billion in future IPO. In the meantime, Galbot has raised more than $300 million in new funding, reportedly pushing its valuation to $3 billion, one of the largest fundings to date in China’s humanoid robotics sector.

American companies are also moving beyond flashy demonstrations to focus on real-world deployments. Moreover, they pursue their own aggressive goals. American Startup Foundationfor example, plans to build 50,000 humanoid robots by the end of 2027.

But China is already targeting a mix of affordable mass-market models and high-end applications, rapidly expanding humanoids in the industrial, consumer and rehabilitation sectors, according to a December study. TrendForce Report.

Obstacles to Chinese domination

When it comes to AI systems and embedded software, it is still unclear where China’s humanoid companies actually stand. The industry is betting heavily on vision-language-action models and “global models,” but both technologies are still in their infancy. Nvidia currently leads the field with its end-to-end humanoid software stack, according to Xu, so naturally most humanoid startups in China are powered by Nvidia’s Orin chips. However, domestic chipmakers are developing local alternatives, she said.

Yet humanoid robot makers are still working on fundamental problems. The challenge is to enable the robot’s core models to predict the “next physical state” the robot will encounter in unpredictable environments, similar to how large language models predict the next word. But unlike LLMs, humanoid robotics companies can’t just grab training data from the Internet, Xu said. Most therefore rely on simulation environments, which generate synthetic data, even if the collection of real data remains essential.

“Because of the data scarcity problem, humanoids are still far from autonomy. The hardware is currently ahead of the software – the robot body can handle much more dexterity today than a few years ago (although it has reliability issues, as we saw with robots crashing during humanoid marathons), but the brain is still in its infancy,” the analyst said.

Safety is also a major obstacle for humanoid robots. A high-profile accident could trigger a public backlash, and China is likely considering how to quickly deploy the technology without moving too quickly. As the industry matures, more regulations are expected.

Given the lack of data, Zhao believes that demand for humanoids will first increase in relatively confined workplaces.

“The first boost will likely be in industrial manufacturing, warehouse logistics and retail, where tasks are repetitive, hours long and processes 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 race between two countries. The Japanese robotics ecosystem – from startups to semiconductor heavyweights – is aim for mass production of humanoids by 2027. Long a pioneer thanks to projects like Honda’s Asimo, Murata Manufacturing’s Murata Boy and SoftBank Robotics’ Pepper, Japan relies on precision and advanced control. An area unique to this country: humanoid robots are increasingly used in elderly care.

James Riney, CEO of Coral Capital, which invests in technology companies in Japan, believes Tokyo will continue to thrive in the humanoid robotics industry. “There are three factors likely to drive the adoption of robotics in Japan. The first is the labor shortage and the desire to rely less on mass immigration. The second is the widely held cultural view of robots as our friends – more like Doraemon versus Terminator. The third is that Japan is already dominant in many parts of the robotics supply chain.”

Hyundai Motor’s Boston Dynamics unit introduced a new humanoid Atlas for factory use by 2028with the intention of producing up to 30,000 units per year in the United States as part of its AI-driven robotics campaign.

Yet for China, government policy, industrial strategy, labor shortages and private capital are all converging to power the country’s rise in humanoid robotics.

“China’s leadership is best understood as an advantage in terms of speed of scaling,” Zhao said. “The ecosystem here compresses the entire cycle – R&D, supply chain, manufacturing, onboarding and customer deployment – into a very tight loop. This means that humanoid companies can go from prototype to real-world deployment more quickly, learn from real-world operations and iterate at a pace that is difficult to match anywhere else.”