Chinese humanoid robot manufacturer Unitree is entering the stock market at a time when investor interest in the industry continues to grow rapidly. The company has finalized preparations for its upcoming IPO on the Shanghai Stock Exchange, setting the offering price at 150.8 yuan per share, implying a valuation of 61 billion yuan ($9.04 billion). The manufacturer intends to raise about $904 million, with AI model developer DeepSeek set to become one of its key investors and strategic partners.
The IPO is coming as the wider humanoid robot market starts to move beyond research and demonstrations. Coherent Market Insights estimates that the global Humanoid Robot Market will be valued at USD 6.40 billion in 2026 and is expected to reach USD 104.95 billion by 2033, growing at a CAGR of 49.0% from 2026 to 2033. Growing use of AI in robots, advances in robotics technology, and the need for assistance as populations age are among the factors supporting this growth.
DeepSeek will invest approximately $20.8 million in Unitree, acquiring 2.31% of the shares offered. However, the nature of the collaboration itself is far more important. The companies are effectively combining expertise in two key areas of modern robotics. Unitree is responsible for the hardware platform, while DeepSeek is expected to help address one of the industry’s biggest challenges — developing an intelligent control system capable of understanding its environment and performing complex physical tasks without predefined scenarios.
It is the software that remains the main barrier to widespread adoption of humanoid robots today. Modern machines can already walk, run, and perform simple actions, but fully autonomous operation requires training on vast amounts of data about the physical world. Partnering with Unitree will allow DeepSeek to accelerate the development of multimodal models that process text, images, and spatial information, while the robot manufacturer, in turn, will gain access to some of China’s most promising AI models.
The combination of better AI and better hardware is one of the main trends in the humanoid robot market. Better AI models can help robots understand their surroundings, follow instructions, recognize objects as well as respond to people. At the same time, improvements in sensors, processors, motors, and other hardware are helping machines move and work more effectively. Another trend is the growing interest in elderly care and personal assistance as aging populations and staff shortages create demand for help with everyday tasks. CMI identifies advances in robotics technology, aging populations, and the need for elderly care among the factors supporting market growth.
Investor interest in such projects is driven not only by expectations of future demand but also by the gradual emergence of real-world commercial use cases. One notable example was an experiment conducted by scientists at the University of California, San Diego, in which surgeons remotely controlled humanoid Unitree G1 robots to perform gallbladder removal procedures on two live pigs.
Widespread practical use of such systems is still a long way off. During the tests, the robots required recalibration, and signal latency remains too high for surgical applications. Nevertheless, the experiment demonstrated the potential to use relatively inexpensive humanoid platforms in areas where traditional surgical systems, costing between $500,000 and several million dollars, are not economically viable. Even the Unitree G1, equipped with specialized manipulators, costs approximately $67,000, which significantly lowers the entry barrier for medical institutions.
Another factor that could strengthen Unitree’s appeal to investors is its expanding partnership with Nvidia. The companies unveiled the Isaac GR00T reference platform, which combines the humanoid robot Unitree H2 with Nvidia’s software ecosystem for AI development and training.
In effect, this is an attempt to create the equivalent of Android for humanoid robots. Startups and research centers will be able to focus on developing their own software without spending years building a hardware platform from scratch. The robot is equipped with a new Nvidia Jetson AGX Thor computing module based on the Blackwell architecture, and the Isaac ecosystem includes tools for simulation, training, remote control, and scaling the resulting applications.
