Unitree Robotics, China's leading humanoid robot maker by shipment volume, just went public on Shanghai's STAR Market, but the company's own IPO prospectus reveals a sobering reality: despite shipping over 5,500 humanoid units in 2025, only 3 to 4 percent are actually performing practical industrial tasks. The rest are deployed for research, university studies, corporate receptions, tour guide duties, and entertainment purposes.

The Hangzhou-based company submitted its IPO prospectus to regulators on March 20, 2026, marking the first time audited financial figures entered the public record for a major Chinese robotics firm. The China Securities Regulatory Commission approved the listing on July 3, 2026, just 73 days after application, one of the fastest reviews on the STAR Market. Unitree planned to raise approximately 4.2 billion yuan, or about 619 million dollars.

What Do Unitree's Financial Numbers Actually Tell Us?

On the surface, Unitree's 2025 results look impressive. Revenue hit 1.708 billion yuan, a 335 percent increase year-over-year, with a gross margin of 60.27 percent. The company reported a net profit of 287.6 million yuan. Humanoid robots accounted for 51.5 percent of revenue in the first nine months of 2025, up from just 1.9 percent in 2023, signaling a dramatic shift in the company's business mix.

But that 60 percent gross margin is the key to understanding what Unitree is actually selling. For context, UBTECH, another listed humanoid robot maker, maintains gross margins around 37 percent. Even Apple, one of the world's most profitable companies, recorded only 48.2 percent gross margin in early 2026. Unitree's profitability far exceeds industry norms, and the prospectus explains why.

Industrial applications account for only 9 percent of Unitree's humanoid sales volume. Of that 9 percent, approximately 50 to 70 percent are used for corporate reception and tour guide purposes, not actual manufacturing or inspection work. This means only 3 to 4 percent of total humanoid shipments are deployed for practical industrial tasks such as inspection and testing. The remaining 90 percent-plus go to research institutions, universities, developers, events, and entertainment venues.

Why Does This Matter for the Robotics Industry?

Unlike industrial machinery delivered to factories, robots destined for research labs, university classrooms, and entertainment events face fundamentally different reliability requirements. They do not need to run 24 hours a day, maintain high uptime rates, or stop production lines if they fail. If a research robot breaks, it can simply be returned for repair. This explains how Unitree maintains such a high profit margin while continuing to lower prices and expand shipments.

The prospectus itself acknowledges that industrial deployment remains limited due to immature technology. This is not a criticism of Unitree's strategy, but rather a rational business sequence: enter a market with lower reliability requirements, build manufacturing scale and experience, generate revenue and profits, then transition to more demanding industrial applications.

The problem, according to industry observers, is that this distinction has been largely lost in public discussion. Shipment numbers have taken on a life of their own, creating an image in global markets that humanoid robots are already working in factories across China. The IPO prospectus directly contradicts this narrative.

How Are Celebrity Investors Betting on Chinese Robotics?

Beyond Unitree's public market debut, the robotics sector has attracted unexpected attention from entertainment celebrities in China, who are shifting investment focus away from the once-trendy hot pot restaurant business toward hard technology ventures. Actor Huang Xiaoming, best known for co-founding the now-defunct Spicy No.1 hot pot chain, has made consecutive investments in robotics companies through his Mingjia Capital fund.

Huang Xiaoming's robotics portfolio includes stakes in Tianzu Zu, a robot leasing platform incubated by Agibot Robotics, and Critical Point, a dexterous hand company spun out from Agibot. Mingjia Capital participated in Tianzu Zu's angel round in March 2026, which raised over 100 million yuan, and followed up with a Pre-A investment one month later. Tianzu Zu is now valued at 7 billion yuan. Critical Point, which completed three consecutive financings within six months, reached a valuation of 1 billion US dollars by May 2026.

Huang Xiaoming's entry into robotics began during a television variety show recording. In a Mango TV program called "Chinese Restaurant: Africa Startup Season," a robot named "Xiao Jiu" served dishes to guests. The robot was jointly developed by Agibot and the media company, based on Agibot's Expedition A2 platform. As a resident guest on the show, Huang Xiaoming formed a connection with Agibot's leadership and later became an active investor and public supporter of the company.

Other entertainment industry figures are making similar moves. Wang Yibo's agency, Yuehua Entertainment, participated in Tianzu Zu's angel round and has discussed combining robots with concerts and fan meetings. Superstar Legend, a concept stock associated with Jay Chou, announced in July 2025 that it would cooperate with Unitree Technology on developing quadruped robots and consumer robotics products with entertainment intellectual property attributes.

Steps for Understanding the Real State of Industrial Robotics Deployment

• Distinguish shipment claims from deployment reality: When a robotics company announces shipment numbers, ask what percentage are actually performing productive work versus serving research, education, or entertainment purposes. Unitree's own prospectus shows this gap is substantial.
• Examine gross margins as a business model indicator: Unusually high profit margins in robotics often signal that units are being sold to low-reliability-requirement markets like research and events, not to demanding industrial customers who require 24/7 uptime and production guarantees.
• Request multi-week pilot data before believing scale claims: Announcements of autonomous warehouses or factory-wide robot deployments should be backed by independent benchmark reports, factory videos, and verified throughput metrics, not just press releases.
• Track which customers are actually ordering robots: Real industrial demand is beginning to emerge in 2026, with companies like State Grid, China's largest power transmission company, placing procurement orders. These real orders are more meaningful than total shipment counts.

What Does the Broader Market Think About Humanoid Robot Companies?

The stock market is sending a mixed signal about the robotics sector. UBTECH, which listed in Hong Kong, saw its humanoid robot revenue grow 2,203.7 percent year-over-year in 2025, reaching 820.6 million yuan. The company's full-year revenue grew 53.3 percent to 2.001 billion yuan, and its net loss narrowed by approximately 32 percent. UBTECH is shipping its industrial Walker S2 robot from a factory in Liuzhou, with orders since early 2025 exceeding 800 million yuan. Customers include major automakers like BYD, Geely, and FAW-Volkswagen.

Despite these achievements, UBTECH's stock price fell 35 percent in 2026 and is near its 52-week low. This divergence reflects investor skepticism about the gap between technical accomplishment and actual business results. Being able to build and ship 1,000 units is different from those 1,000 units actually increasing customer productivity and generating return on investment.

AgiBot, an unlisted competitor, reportedly led all rivals with 5,168 humanoid shipments in 2025 and reached 10,000 cumulative units by March 2026. The company is preparing for a Hong Kong listing with a target valuation of approximately 800 billion to 1 trillion Japanese yen. Shareholders include Tencent, BYD, and other major investors. In April 2026, AgiBot set a goal of 10 billion yuan in sales within five years and 100 billion yuan within eight years.

The question of which company leads in shipment volume remains murky. Unitree claims over 5,500 units and a 32.4 percent market share in its prospectus, while research firm Omdia states that AgiBot's 5,168 units led all competitors. The discrepancy likely stems from different counting methods, such as whether companies are counting shipments, production runs, or units that have left the factory, and whether quadruped robots are included in the tally.

For Indian developers and builders interested in open-source robotics, the landscape is shifting toward accessible hardware and standardized software stacks. The Unitree G1 and H1 humanoid robots ship with ROS 2 control stacks, joint torque sensors, and published URDF files, making them viable platforms for research and development. Other affordable entry points include the Dobot Magician and Elephant Robotics myCobot 280, priced between approximately 35,000 and 55,000 Indian rupees.

The robotics ecosystem is maturing, but the gap between research-grade deployments and industrial-scale production remains wide. Unitree's IPO prospectus has, for the first time, provided audited evidence of this reality. The company is not yet selling labor at scale; it is selling tools for development and devices to attract attention. Whether that changes in the next few years will determine whether the current wave of robotics investment delivers on its promises or becomes another tech bubble.