Unitree's $9B IPO Tests Whether China's Dancing Robots Can Earn Real Money

Unitree Robotics priced its Shanghai IPO at 150.8 yuan per share—about $22.40—on a Thursday in August, raising $900 million and landing a $9 billion valuation. The Hangzhou startup became the first humanoid robot maker to list on mainland China's STAR market, and the market responded with something close to delirium. Retail investors oversubscribed the online tranche more than 5,000 times. The winning lottery rate hit 0.018%, a historic low. On crypto derivatives exchanges, pre-IPO contracts on Unitree were trading at roughly four times the IPO price, a sign that money was chasing the story faster than the company could tell it.

But beneath the frenzy sat a harder question: What are these robots actually for? Unitree's machines have become famous for their acrobatics—they can throw kung fu kicks, bounce back from knockdowns, dance across stages. They are, by any measure, impressive feats of engineering. Yet when Hao Hong, managing partner of Lotus Asset Management, looked at what Unitree's robots could do in the real world, he saw a gap. "For these humanoid robots, to be honest, they're fascinating. They can dance and all that – but never seen them doing any real housework," he said. The company's own prospectus acknowledged the constraint: robotic hands are still not precise or durable enough for sustained, practical work. Dominik Pross, an equity analyst at VP Bank, added that even advanced humanoids can typically handle only a small number of tasks, and only for a few hours before their batteries demand a recharge. Most models run for up to four hours while sitting idle. Each robot must be specifically trained for each task, even simple ones. Complex everyday activities remain out of reach.

Unitree's revenue more than quadrupled last year, a fact that lent credibility to the bull case. But the company's first-quarter adjusted profit fell more than 52 percent as spending on research, development, and marketing climbed. Nearly three-quarters of its humanoid revenue came from research and education in the first nine months of 2025. Corporate tours and demonstrations made up more than half of its still-small industrial business. The core problem was not growth but proof: investors were betting on a future that had not yet arrived.

The economics, however, were moving in Unitree's favor. Average humanoid robot prices have plunged 93 percent between 2020 and 2025, falling to $58,000. Unitree's flagship G1 model costs just $82 a year in electricity to run for eight hours daily. The company's cheaper G1 EDU model, built for research and customization, had been cut in price by more than 45 percent in a single year to $27,300, while the company maintained a 67 percent gross margin. Lower-cost manufacturing had given China a decisive edge. The country now accounts for more than 70 percent of global industrial robot installations and nearly 90 percent of all humanoids deployed. Wood Mackenzie projects the global humanoid robot fleet will grow more than 90 percent annually through 2035, surpassing 10 million units, with annual shipments topping 4 million by then. Chinese government support—including investments in domestic manufacturers' shift from labor-intensive work to automation—was fueling the momentum.

Yet geopolitical headwinds were gathering. The U.S. moved to ban imports of foreign-made humanoid and four-legged robots last month, a move widely understood as targeting Chinese technology. Unitree derived 13 percent of its revenue from the U.S. last year, making it particularly exposed. Some U.S. university researchers said they were unfazed because small-batch imports for testing and development remain exempt. But Unitree's four-legged robots had also been deployed in U.S. state correctional systems and military settings, according to the House Select Committee on China, raising national security concerns that went beyond academic interest.

Another risk loomed larger: the possibility that Chinese robotics companies could lose access to Nvidia's hardware and software stack. China's industry leaders were among the early users of Nvidia's platforms for powering robots, and they were not yet positioned to operate without Western components entirely. But China held a countervailing advantage. The country's dominance in rare earths—essential materials for the actuators and motors that make humanoids move—gave Chinese players significant leverage over global competitors. Meanwhile, Tesla CEO Elon Musk was doubling down on the sector, converting EV production lines at the company's Fremont factory to manufacture Optimus humanoid robots, with production expected to start later in the year.

Unitree's $9 billion valuation—over 200 times last year's earnings—reflected not just the company's current performance but a bet on a market that barely existed yet. Rival companies AgiBot and Leju Robotics were preparing their own Hong Kong and Shenzhen listings. The humanoid robotics industry had a long way to go, but the potential market opportunity had convinced investors, venture capitalists, and major manufacturers that the moment had arrived. Whether the robots themselves were ready was a different question entirely.