Unitree Robotics has become a symbol of China’s push into humanoid robots after a striking Shanghai market debut. The Hangzhou company raised 6.1 billion yuan ($904 million), and its closing price valued the business at around $50 billion.
The rally was dramatic: Unitree’s stock jumped as much as 629 percent during its Shanghai debut and closed up 460 percent, according to Bloomberg. That made it the first publicly traded maker of humanoid robots on the Chinese mainland.
A Fast Rise With A Complicated Demand Story
The central question is not whether investors are excited about robots. It is what kind of demand sits underneath that excitement.
According to the Financial Times, Chinese humanoid robot makers sell many machines to state-backed training centers. Those centers use people to operate the robots remotely and teach them physical tasks. The centers then sell the resulting training data back to the manufacturers.
That structure can create activity on both sides of the market. Robot makers record sales of machines. Training centers collect data that manufacturers need for development. Local governments and manufacturers often fund the centers together.
Analyst Poe Zhao summarized the concern directly: "It also blurs the distinction between independent demand and demand created within a policy-supported ecosystem."
How The Training Center Loop Works
The model is built around a practical problem for humanoid robots: they need training data for physical behavior. A robot that can move, dance, handle objects or perform tasks depends on examples gathered through use.
In this system, the buyer is often not an ordinary commercial customer using robots in everyday operations. Instead, the robot may go to a training center where human operators guide it through tasks by teleoperation. The data produced through that process can then be packaged and sold back to the same industry that supplied the machines.
By June there were more than 90 such centers, according to Interact Analysis. The price of the data can be high: training data for a five-minute robot dance can cost up to one million yuan ($148,000).
The sales exposure is meaningful. At manufacturer Leju, training centers accounted for 45 percent of sales of its flagship robot. At Unitree, nearly three-quarters of humanoid revenue in the first nine months of 2025 came from education and research.
One algorithms engineer described the incentives this way: "Both sides get what they need, and both sides have something to show."
Why Analysts Are Skeptical
The first concern is valuation. Unitree was valued at 35.89 times revenue, according to Bloomberg. Its Hong Kong rivals were valued at about 20 times revenue.
Vey-Sern Ling of Union Bancaire Privée was blunt about the market reaction: "There's clearly no fundamental basis for the share price surge." The Financial Times also reported that early-stage investors are already looking for an exit.
The second concern is whether the training data itself is as valuable as the spending suggests. Marco Wang of Interact Analysis said the data is not "100 percent useful" because the robots are not operating in real-world settings. One center manager said only two to three of every eight training hours are usable.
That matters because the circular financing issue is not only about who buys the robots. It is also about whether the products, data and end demand are moving toward durable commercial use or staying inside a policy-supported ecosystem.
A Broader AI Pattern
The Unitree case echoes a debate already visible in the US AI industry. The source article compares the setup with Nvidia’s direct and indirect investments in AI firms, including in Ohio, where Nvidia is backing a new OpenAI data center as a guarantor.
Critics see these arrangements as a way for demand to reinforce itself and make the market look larger than it is. Nvidia CEO Jensen Huang has argued that these and other investments are needed to support fast growth by a company that cannot build the foundations for that growth entirely on its own.
Beijing appears to view the robotics sector through a similar growth lens. About 20 percent of Unitree’s IPO went to strategic investors, including AI startup Deepseek. Supporters point to electric cars and solar panels, sectors China now dominates after state-driven demand helped them get started.
What To Watch Next
The debate around Unitree is ultimately about signal quality. A $50 billion valuation, a 460 percent closing gain and a large IPO all point to investor enthusiasm. But the structure of demand determines what those numbers mean.
If robot purchases are driven mainly by state-backed centers that sell data back into the same ecosystem, then revenue may say less about broad market adoption than it first appears. If the data becomes useful enough to improve robots and support wider demand, the system could still become a foundation for growth.
For now, the facts point in both directions. Unitree has public-market momentum and a policy-backed environment. It also faces questions about circular demand, data usefulness and whether its valuation has moved faster than the commercial case for humanoid robots.