The flood of venture capital pouring into Chinese robotics companies has officially hit its first major reality check.
Unitree Robotics just cleared its final regulatory hurdle from the China Securities Regulatory Commission (CSRC) for a massive initial public offering on Shanghai's tech-heavy STAR Market. The company wants to raise roughly 4.2 billion yuan ($619.4 million), aiming for a valuation that could hit anywhere between $6 billion and $7 billion depending on final pricing. Also making news lately: Why Meta Cannot Program Its Way Out Of The Eus Latest Crackdown.
It took just 104 days from filing to regulatory sign-off. That is blindingly fast for China's market. But do not let the speed fool you. This isn't just another tech company hitting the public markets. It is a high-stakes stress test for an entire industry that has relied on hype, massive government targets, and desperate venture capital looking for the next big AI play.
Everyone is watching to see if public investors will actually buy what the private markets have been hyping up. Additional details on this are covered by MIT Technology Review.
The Margin Illusion Behind the Hype
If you look closely at Unitree's prospectus, you notice a weird contradiction. The average selling price of their humanoid robots dropped from around $85,000 in 2023 to just $25,000 in late 2025. In any standard hardware business, a pricing crash like that signals a race to the bottom that destroys profits.
Yet, Unitree's companywide gross margin climbed to nearly 60%.
How? The answer comes down to manufacturing leverage. McKinsey notes that actuation and core physical components eat up about 40% to 60% of a humanoid robot's bill of materials. Unitree builds its own components rather than buying them off the shelf. By scaling up internal production, they are driving their internal costs down faster than they are cutting prices for buyers.
Humanoids are no longer a side project for this team. In 2023, humanoids made up less than 2% of Unitree's core revenue. By late 2025, they brought in over 51%, overtaking their famous four-legged robot dogs. They produced 3,701 humanoids in the first nine months of 2025 and sold 3,551 of them. That is a 95.95% sell-through rate.
That tells us people are buying. It does not tell us who is buying or why.
The Software Trap That Most Investors Ignore
There is a massive gap between a robot that looks great in a choreographed video and a robot that can actually do useful work in a factory or home.
China's robotics market is heavily crowded. There are roughly 150 companies chasing the exact same market right now. Rivals like AgiBot and EngineAI are aggressively eyeing Hong Kong listings, while smaller names like Dobot and Deep Robotics wait in the wings. EngineAI even explicitly used Unitree’s private valuation as a benchmark in its own filings.
But here is what most people get wrong. The entire sector is bloated with hardware announcements and desperately thin on software.
Building a shiny metal body that can walk is a solved problem. Building the "brain"—the AI models and control systems that allow a machine to adapt to an unpredictable environment—is where the real struggle lies. Unitree is being open about this. Their prospectus explicitly states that the IPO proceeds will go toward robot artificial intelligence model research, alongside building out a smart manufacturing base.
They know the hardware alone will not save them. If a humanoid robot cannot think or adapt, it is just an expensive piece of industrial sculpture.
Beijing Pressure Meets Geopolitical Reality
You cannot talk about Chinese robotics without talking about the government. Beijing has set a rigid national target to deploy 100,000 humanoid robots by 2027 under an aggressive action plan backed by five separate ministries.
This political backing is why the domestic market has exploded. In fact, mainland China’s share of Unitree’s revenue jumped to over 60% recently, up from less than half in previous years.
But this heavy domestic reliance highlights a growing vulnerability. Unitree’s own filings flag significant tariff and export-control risks. The company has already found itself on a Pentagon list of Chinese companies with alleged military ties. Relying on China's domestic market works great while the government is subsidizing the push, but it limits the ultimate ceiling for global growth.
Public markets are famously less forgiving than venture capitalists. Private investors can ride a wave of paper valuations for years based on expectations. Public equity markets demand consistent earnings, transparent data, and a clear path to global scale.
Next Steps for Tech Investors
If you are tracking the robotics space or looking to invest, do not get blinded by the $6 billion headline numbers. Here is how you should evaluate the market moving forward.
- Watch the final pricing: The gap between Unitree's final listing price and its target valuation will tell you exactly how much skepticism public investors have about the humanoid hype cycle.
- Audit the buyer profile: Look closely at future earnings reports to see if these robots are going to real enterprise buyers using them for labor, or if they are just being bought by state-backed institutes to fill quotas.
- Track software milestones: The companies that win will not be the ones that build the cheapest bodies, but the ones that develop the most adaptable software brains. Keep your eyes on Unitree's actual R&D spend on AI models over the next twelve months.
The era of easy venture capital for any company with a cool walking robot is ending. Unitree is about to show everyone whether this industry is built on a real foundation or just really expensive science projects.
