Key Takeaways
- Unitree’s shares have surged by around 500%, pushing the company’s valuation to roughly $62 billion.
- Its humanoid robot, known as “Superman,” can reach speeds of up to 28 mph, while China now has more than 330 humanoid robot models in development.
- Despite its rapid growth, Unitree is facing increased scrutiny in the U.S. after being flagged by the Pentagon in June 2026.
Unitree hit the Shanghai Stock Exchange and immediately looked less like a hardware startup and more like a market event, with shares jumping roughly 500% and briefly running even hotter at the open. The rally followed fresh footage of its “Superman” humanoid sprinting at 12.66 meters per second and leaping two meters, a neat party trick that also signals just how fast China’s robot supply chain is tightening up. By mid-morning, the surge implied a valuation of about $62 billion, putting a Hangzhou robot maker that counts Nvidia, Google, and Amazon among users into the global spotlight. It is also landing in the middle of Washington’s tech clampdown, where the Pentagon has flagged the company and new US rules are moving to curb imports of foreign humanoid robots.
Unitree’s loud arrival on the public stage
If you’ve been watching robotics from afar, this week offered a jolt of clarity: investors are ready to pay up for physical AI. Unitree, the Hangzhou-based humanoid robotics maker, made its trading debut on the Shanghai Stock Exchange and immediately became one of the most talked-about tech listings of the year, even from across the Pacific.
The stock opened near $21 per share and quickly surged, changing hands around $123 later in the morning, roughly a 486% jump from its IPO price. At those levels, Unitree’s market value was about $62 billion, an eye-popping figure for a company still better known for demos than for mass deployment.
A robot named “Superman,” and a message to the market
The timing was not subtle. Just days before the listing, Unitree unveiled a new humanoid model called “Superman”, and the specs were built for headlines: the company says it can run about 28 mph and jump roughly 6.5 feet. That kind of athletic performance is catnip for social media, but it also signals how intensely these firms are tuning hardware for real-world mobility.
Unitree’s machines also show up in the orbit of American tech. People familiar with its customer base say its robots have been used by US giants including Nvidia, Google, and Amazon, often as platforms for research and prototyping. For US readers, that matters because it hints at an uncomfortable truth: even when the hottest AI models are trained in the US, plenty of the embodied experimentation is global.
Robotics at scale, and a supply chain advantage
China has treated robotics as a national priority, and the numbers are hard to ignore. Chinese officials have said more than 140 domestic companies introduced over 330 humanoid robot models in recent years. That kind of breadth creates a fast feedback loop between parts suppliers, factories, and labs, the same dynamic that once accelerated smartphones and drones.
Unitree has benefited from that environment, becoming a visible “face” of the category. The IPO pop, in that sense, was not just about one firm’s balance sheet. It was a bet that humanoids are moving from showcase stunts to a durable product cycle.
US scrutiny rises as physical AI gets real
Here’s the catch for American businesses: Washington’s stance is tightening. In June, the Pentagon added Unitree to a list of entities described as “Chinese military companies” operating directly or indirectly in the US, per the government disclosure. In July, the US government also moved to restrict imports of foreign-made humanoid robots on national security grounds, a step that would fall hardest on Chinese suppliers.
So what happens next: a cross-border tech story, or two parallel ecosystems? Unitree’s debut suggests investors think there’s plenty of demand either way, but the next chapter will hinge on whether these machines can prove themselves outside the lab, under real operational constraints.

