Key takeaways
– Public subscriptions for Unitree Robotics’ Shanghai STAR Market offering were scheduled to begin on August 10, 2026.
– The company priced 40.45 million new shares at 150.8 yuan each, implying proceeds of about 6.1 billion yuan and a post-offer valuation near 61 billion yuan.
– The listing matters because it gives public markets a rare, relatively transparent benchmark for a profitable humanoid-robot manufacturer.
– Demonstrations are not the same as dependable workplace autonomy. Data, safety, maintenance, software reliability and useful deployment economics remain decisive.
Unitree Robotics reached a significant commercial milestone on August 10, 2026, when public subscriptions were scheduled to open for its initial public offering on Shanghai’s technology-focused STAR Market. The offering turns one of the world’s most visible makers of legged robots into a public-market test of whether humanoid robotics can move from impressive demonstrations to durable businesses.
The Hangzhou company priced the offer at 150.8 yuan per share. It plans to issue roughly 40.45 million new shares, equal to 10% of its enlarged share capital, and raise about 6.1 billion yuan. Reuters calculated that the price implies a valuation of approximately 61 billion yuan, or about $9 billion at the exchange rate used in its report.
Those figures are important, but the larger story is industrial. Unitree has built a recognizable brand around relatively affordable quadruped and humanoid machines. Its listing creates a public reference point for the cost, growth and risks of “embodied AI”—systems that combine artificial intelligence with machines able to sense and act in the physical world.
What happened on August 10
The August 10 subscription date was set in Unitree’s offering schedule after the Shanghai Stock Exchange accepted its application, reviewed the filing and approved the listing process. The offer price was established on August 6.
Unitree is expected to become the first mainland-listed Chinese company primarily identified with humanoid robots. That description needs care: other public companies manufacture robots or automation systems, and Hong Kong-listed UBTech also develops humanoids. The milestone is specifically about a mainland A-share listing and Unitree’s positioning as a specialist in humanoid and quadruped robots.
DeepSeek participated in the strategic placement. Reuters reported that the AI company invested 140.8 million yuan for 933,399 shares, representing 2.31% of the offering rather than 2.31% of Unitree’s entire post-offer equity. That distinction prevents a common misreading of the placement.
What the filing says about Unitree’s business
Unitree began with four-legged machines and later expanded into humanoids. Its products now include the Go and B-series quadrupeds as well as G1, H1 and R1 humanoid platforms. Videos of the machines running, dancing and performing coordinated movements helped the company become widely recognized outside industrial robotics circles.
According to figures reported from the prospectus, Unitree generated about 1.7 billion yuan in revenue in 2025, more than four times the previous year’s level. Humanoid products produced around 867.8 million yuan, overtaking quadruped robots as the company’s largest reported product category.
The filing also shows why growth headlines should not be read without context. First-quarter 2026 revenue continued to rise, but adjusted profit declined as research, development and marketing costs increased. This is a normal tension in an emerging hardware category: companies must fund engineering, manufacturing and distribution before there is certainty about repeatable demand.
Overseas sales represented more than 40% of 2025 revenue, according to Reuters’ review. That gives Unitree global reach, but it also creates exposure to tariffs, export controls, procurement restrictions and changing product approvals.
Where the IPO money is expected to go
Unitree’s earlier draft prospectus sought 4.2 billion yuan for core technology and manufacturing projects. The final offer price increased expected gross proceeds to about 6.1 billion yuan. Company disclosures and reporting identify three broad priorities:
1. Improving robot hardware and AI models.
2. Developing new products and supporting technologies.
3. Expanding intelligent manufacturing capacity.
These priorities illustrate why robotics requires more than a strong language model. A useful humanoid must combine perception, planning, balance, manipulation, battery management, thermal control, connectivity and safe interaction. It must then be manufactured consistently and supported in the field.
Why the listing matters beyond China
It creates a public robotics benchmark
Many prominent humanoid programs sit inside large companies or private startups. Their costs, shipments and profitability are difficult to isolate. Unitree’s public reports will make revenue mix, margins, research spending and risk disclosures easier to examine over time.
That does not make Unitree a perfect proxy for every robotics company. Its product mix and Chinese supply chain are distinctive. It nevertheless gives analysts and competitors a more concrete reference than promotional videos or private funding valuations.
It tests demand outside controlled demonstrations
Modern humanoids can perform visually impressive sequences. Commercial customers need a different standard: Can a robot perform a useful task repeatedly, in a changing environment, at an acceptable total cost and without creating safety problems?
Education, research, inspection and guided demonstrations are already plausible markets. Unsupervised work in homes or complex factories is harder. A listing will expose the gap between shipped units, revenue and reliable real-world deployment more clearly.
It connects AI competition to manufacturing
The strategic placement involving DeepSeek highlights an important direction. Language and vision models need physical data and control systems before they can operate machines effectively. Robot makers need better models, while AI developers need real-world interaction data.
The partnership could support that exchange, but it does not prove that general-purpose autonomy has been solved. Physical errors can damage equipment or injure people; the reliability threshold is much higher than for a chatbot response.
Original analysis: the four gates between a robot demo and a scalable business
The Unitree listing can be assessed through four practical gates.
| Gate | Question | Evidence that matters |
|---|---|---|
| Capability | Can the robot complete the target task? | Success rate across varied environments, not one edited demonstration |
| Reliability | Can it repeat the task safely? | Uptime, failure modes, intervention rate and safety testing |
| Economics | Is it cheaper or more useful than alternatives? | Total ownership cost, maintenance, operator time and productive hours |
| Deployment | Can customers integrate and support it? | Training, software interfaces, spare parts, service coverage and cybersecurity |
A robot can pass the capability gate and still fail commercially. For example, a warehouse machine that occasionally carries a box may not be valuable if a worker must reset it after every few attempts. The practical metric is useful autonomous work per unit of total cost, not the most dramatic movement shown on video.
Risks readers should keep in view
Unitree’s overseas exposure makes trade restrictions a material risk. Reuters reported that existing models had received relevant U.S. approvals while future products could face restrictions. Any statement about market access therefore needs a date and product-specific context.
Competition is another risk. Chinese robotics companies are moving quickly, and established automation firms already have customer relationships and service networks. Price reductions can increase adoption but also compress margins.
Technical risk remains fundamental. Humanoid machines need better dexterity, autonomy, battery life and reliability. Cybersecurity is also essential because a connected robot combines cameras, microphones, motors and network access. Buyers should demand documented update policies, access controls, vulnerability handling and data-governance terms.
Finally, an IPO price is not proof of long-term commercial success. It reflects the terms of one capital-raising event and current demand for the shares. The more useful signals will arrive later through customer concentration, repeat orders, service revenue, warranty costs and research productivity.
What to watch next
The first item is completion of the offering and the start of public trading. After listing, early price movement will attract attention, but operational disclosures will be more valuable.
Watch the share of revenue generated by humanoids, the balance between domestic and overseas demand, adjusted profit, research spending and manufacturing expansion. Also watch whether customers describe repeatable deployments rather than trials.
Product announcements should be evaluated against the four gates above. A lower price is meaningful only if reliability and support remain strong. A more capable model matters only if customers can integrate it safely.
Frequently asked questions
When did Unitree’s IPO subscriptions open?
The published offering schedule set August 10, 2026, for public subscriptions on Shanghai’s STAR Market.
How much is Unitree raising?
At 150.8 yuan for roughly 40.45 million new shares, expected gross proceeds are about 6.1 billion yuan before offering expenses.
Is Unitree the first publicly listed humanoid-robot company?
It is expected to be the first specialist humanoid-robot maker listed on mainland China’s A-share market. Other robotics businesses, including UBTech in Hong Kong, are already publicly traded.
Does the IPO prove humanoid robots are ready for mass adoption?
No. It shows investor demand and gives Unitree additional capital. Commercial readiness still depends on reliability, safety, useful autonomy and deployment economics.
