Unitree Robotics Prospectus Revealed: World's No. 1 in Humanoid Robot Shipments, Commercialization Relies on Schools and Research Institutions

Unitree is about to IPO on the Shanghai Stock Exchange's STAR Market, raising $620 million. In 2025, humanoid robot shipments reached 5,500 units, ranking first globally. However, 74% were sold to academic institutions, with actual industrial applications accounting for only 9%, revealing the industry's reality: hardware has been figured out, but AI models still need to catch up. (Previously: TSMC "rumored to cut dividends" — netizens slam executives taking theirs first, calling for #PhoneOffAfterWork movement) (Background: F2Pool co-founder Chun Wang to become SpaceX manned flight commander, explaining "why choose Mars, not the Moon") Unitree Robotics is about to land on the Shanghai Stock Exchange's STAR Market, planning to raise $620 million. This prospectus reveals for the first time the complete financial details of a profitable robotics company: humanoid robot shipments reached 5,500 units in 2025, ranking first globally, but 74% were sold to university research institutions, with actual industrial applications accounting for only 9%. This picture reveals the current truth of the industry — hardware technology has been figured out, but commercialization is still waiting for AI model capabilities to catch up. DeepFlow Brief: Unitree robotics is about to IPO on the Shanghai Stock Exchange's STAR Market, raising $620 million. The prospectus discloses for the first time the complete financial data of a profitable robotics company. The company shipped 5,500 humanoid robots in 2025, ranking first globally, but 74% were sold to universities for research, with real industrial scenarios accounting for only 9% — this is the truth of the current robotics industry: hardware has been figured out, but commercialization is still waiting for AI models to catch up. Unitree Robotics recently submitted an IPO application to the Shanghai Stock Exchange's STAR Market, planning to raise $620 million. This prospectus is quite interesting because it lets us clearly see the real state of the current robotics market. Unitree is already profitable, growing rapidly, and ranks first globally in humanoid robot shipments. Unitree was founded in 2016 in Hangzhou. Founder 王興興 is a self-taught robotics expert who built his first quadruped robot in his own apartment. The company now has 480 employees, of which about 175 are in R&D. The company sells two product lines: Quadruped robots: Go2 (consumer-grade and research-grade), B2 (industrial-grade), and A2 Humanoid robots: H1, H2, G1, and R1. G1 is the one you may have seen in viral videos, standing 1.32 meters tall and weighing 35 kg The company began international sales in 2018. Over 35% of revenue comes from outside China, including a large number of US academic customers. Two years ago, Unitree was essentially a robot dog company, mainly selling quadruped robots. In 2023, humanoid robots accounted for only 1.9% of revenue. By the first three quarters of 2025, humanoid robots accounted for more than half of core revenue. What drove this shift was improved product-market fit and aggressive marketing strategy. The company's humanoid robots appeared on CCTV's Spring Festival Gala for two consecutive years, one of the highest-rated programs globally. Jensen Huang brought a Unitree robot on stage at the 2024 GTC conference. This brand exposure translated into commercial and research demand, something most Chinese hardware companies have never truly achieved. The humanoid robot shipment data is particularly impressive. Unitree shipped approximately 5,500 humanoid robots in 2025, making it the world's largest humanoid robot manufacturer by shipment volume. China's AgiBot comes closest to this figure. By comparison, well-known US companies like Figure AI and Agility Robotics have only shipped a few hundred units. The 5-year target in the prospectus is annual production of 75,000 humanoid robots and 115,000 quadruped robots. That's about 14 times the 2025 humanoid robot shipments. The target is aggressive, but it also highlights how early-stage we still are. The prospectus categorizes buyers into three groups: research and education, commercial consumer, and industrial applications. The reality is that most current humanoid robot demand comes from research and education uses. 1/ Research and education account for 74% of humanoid robot revenue/shipments. Academic buyers have been Unitree's core support since at least 2022 and remain the company's largest source of revenue. 2/ Commercial consumer accounts for 17% of humanoid robot shipments. Non-academic consumers who buy these robots mainly use them for "display": serving as eye-catching promoters in retail venues, tourist attractions, performances, and exhibitions. Consumer revenue in the first nine months of 2025 grew nearly fourfold year-on-year, which sounds impressive, but the base was actually small. In reality, the true use of this $25,000 humanoid robot today is to stand in front of some store in Shenzhen to attract tourists. 3/ Industrial applications account for only 9% of humanoid robot shipments. Unitree acknowledges that industrial deployment is more limited because the technology isn't mature enough, which speaks to the current technological state. Of this 9% of shipments, about 50-70% are used in scenarios like corporate reception and guided tours, so overall only 3-4% of humanoid robot shipments are truly used for work such as corporate reception and inspection. For quadruped robots, the situation is better: only about one-third of revenue comes from research, more than 40% from commercial use, and the rest from industrial use. Productive use cases there are relatively mature. Customers include State Grid, China Southern Power Grid, CNPC, Sinopec, Baowu Group, and JD.com (Unitree's largest customer). These companies use quadruped robots for real inspection scenarios in chemical plants, substations, coal mines, pipelines, and more. A unique aspect of Unitree is that it autonomously designs and manufactures most of the key components: high-torque motors, precision reducers, encoders, joint modules, smart controllers, high-precision sensors, dexterous hands, LiDAR, and cameras. According to McKinsey data, the actuation system (motors, reducers, and the joint systems that really make robots move) typically accounts for 40-60% of a humanoid robot's total bill of materials cost. Most companies in this field source these components externally, but Unitree manufactures them itself. Externally sourced components account for only about 14-18% of total costs. What it outsources are only generic components like batteries and flash memory, as well as differentiated parts such as core computing boards. The unit manufacturing cost of quadruped robots dropped from about $3,300 in 2022 to about $1,800 by mid-2025, a 46% decrease. Humanoid robot costs also declined, from about $10,800 to $9,200 over the same period. Interestingly, as shown in the chart below, the average selling prices of both quadruped and humanoid robots also dropped significantly each year. But gross margins expanded throughout the period, rising from over 40% in 2022-2023 to nearly 60% in 2025, largely due to their vertical integration strategy. Revenue grew from $58 million in 2024 to an estimated approximately $252 million in 2025, an increase of 335%, mainly driven by strong performance in humanoid robots. For most of the company's history, international sales accounted for over 55% of revenue. In 2025, the Chinese domestic market surpassed exports for the first time, although the absolute value of export revenue still more than doubled year-on-year. Gross margins are close to 60% and have continued to expand over the years. For comparison: most hardware companies have gross margins of 30-40%. Software companies typically reach 70-80%. For a company selling physical robots, Unitree's gross margins are relatively high, thanks to their vertical integration strategy and currently relatively differentiated products. The company became profitable under GAAP standards in 2024, with a profit margin of approximately 18%, and close to 35% on an adjusted basis. Unitree's IPO targets a valuation of approximately $6-7 billion. Unitree plans to use nearly half of the IPO proceeds for software. Of the $620 million raised, approximately $300 million will be used over the next three years for AI model training, with about $100 million annually invested in what the company calls "embodied large models." The prospectus describes two parallel model architectures. The first is VLA (Vision-Language-Action): a model that maps directly from visual and language inputs to motor commands, enabling robots to generalize across unfamiliar tasks without hand-coded instructions. The second is WMA (World Model + Action), which is their more favored direction. The WMA model builds an internal simulation of physical reality. The robot predicts what will happen before acting, rather than learning purely through trial and error. They have released initial versions of both. UnifoLM-WMA-0 was open-sourced in September 2025; UnifoLM-VLA-0 was released in January 2026. They have also detailed the approximate spending allocation on the model side, as shown below: Unitree's current hardware lead is real, but the company understands that lasting advantage in robotics may require controlling the model layer: the systems that determine what robots do and how they move. The software ambition is also a hedge against commoditization. Unitree has built a moat in hardware manufacturing. But if actuators and joint modules eventually become standardized components like batteries in electric vehicles, then defensibility shifts to the model layer. Unitree has a profitable hardware business, a real manufacturing moat, and more humanoid robot shipments than anyone else, at prices others cannot reach. But as the actual usage of humanoid robots shows, the broad commercial application story is still in early stages. "Display" use cases dominate consumer demand, and industrial deployment is narrow in scope. Unitree shows us the current state of the robotics market, with much more work to be done on models, hardware, and use cases.