IPO registration takes effect as Unitree moves toward the STAR Market
On July 6, the Shanghai Stock Exchange website updated Unitree Technology’s IPO review status to “registration effective”. Earlier, on July 2, the China Securities Regulatory Commission approved the company’s application for an initial public offering and a listing on the STAR Market. From acceptance of the filing on March 20 to registration taking effect, the process took 104 days, while the exchange review itself was completed in 73 days, a relatively fast timeline under the STAR Market’s pre-review framework.
According to public information cited in the report, Unitree plans to raise 4.202 billion yuan and issue no less than 10% of its shares. Based on prior market expectations, the company’s valuation after listing could reach around 50 billion yuan. For the domestic equity market, this is more than another IPO case: it gives investors one of the clearest pure-play humanoid and robotics assembly stories currently available in A-shares.
Investor attention is tied to Unitree’s unusual positioning. The Hangzhou-based company first built its reputation in quadruped robots, then rapidly pushed into humanoids. After UBTech, it is widely seen as one of the earliest companies in the current humanoid wave to secure a public-market ticket, making its post-listing capital allocation and growth trajectory especially important for the sector.
From a student-built quadruped prototype to a full humanoid lineup
Unitree’s product story starts well before the IPO. In 2013, founder Wang Xingxing, then an undergraduate at Zhejiang Sci-Tech University, built a quadruped robot prototype called XDog during summer lab work. Motors, drivers and the main control board were all assembled by hand. Three years later, he used that prototype and 2 million yuan in angel financing to establish Unitree in Hangzhou.
The company first drew broad industry attention with Laikago in 2017, described in the report as China’s first publicly retailed high-performance quadruped robot. That was followed by AlienGo in 2019 and A1 in 2020, as Unitree kept iterating products while driving prices lower. In 2021, 24 A1 robots appeared on China Central Television’s Spring Festival Gala in a performance with Andy Lau, pushing the company beyond the robotics niche and into public awareness.
The pace accelerated further in 2023. In July that year, Unitree launched the Go2 quadruped with a starting price of 9,997 yuan. Just one month later, it unveiled the full-size general-purpose humanoid robot H1, standing 1.8 meters tall and weighing 47 kilograms. That sequence marked Unitree’s formal move from a quadruped specialist into the deeper waters of humanoid robotics.
The report describes an aggressive release cadence after that. In May 2024, the G1 humanoid intelligence platform launched with a starting price of 99,000 yuan. By June 2026, the company had introduced H2 Plus, equipped with Nvidia’s Jetson Thor and the Isaac GR00T platform. Over roughly two years, Unitree assembled a product matrix stretching from consumer-oriented and developer-facing robots to more industrial-grade systems, spanning from a 26,900-yuan dual-arm humanoid robot to the 3.9 million-yuan GD01 manned mech platform. In between, the company also gained visibility through the 2025 Spring Festival Gala collaboration “YangBot” and the launch of the ultra-lightweight R1 humanoid at 29,900 yuan in July that same year.
Commercial traction is the backbone of the listing story
What differentiates Unitree from many robotics startups is the combination of product volume and actual revenue. The article says that by 2023, the company’s global quadruped robot shipment share had exceeded 60%, with overseas revenue accounting for half of total revenue at that stage. In other words, Unitree was not just demonstrating hardware; it was already selling it at scale and turning those shipments into cash flow.
The prospectus numbers highlighted in the article underscore that point. Revenue rose from 159 million yuan in 2023 to 1.71 billion yuan in 2025, representing growth of more than tenfold in two years. Net profit moved from a loss in 2023 to 288 million yuan in 2025. During 2025, shipments of humanoid robots exceeded 5,500 units, while cumulative sales of quadruped robots surpassed 30,000 units.
Pricing has been one of Unitree’s most powerful competitive tools, but not the only one. A Go2 starting below 10,000 yuan and an R1 starting below 30,000 yuan stand in sharp contrast with an industry that, only two years earlier, often priced humanoids in the six- or even seven-digit range. Overseas markets are also becoming more important. The report says roughly 40% of revenue now comes from abroad. Media coverage cited in the piece adds that, starting in May this year, Japan Airlines began trial operations of Unitree’s G1 at Tokyo Haneda Airport, using the robot for baggage handling, cargo transfer and conveyor coordination, with the test period planned to continue through 2028.
How Unitree spends the 4.202 billion yuan may define its next growth phase
Unitree has outlined several uses for the IPO proceeds: strengthening robot model research, embodied hardware R&D, new product development and the construction of a new manufacturing base. In practical terms, that means spending on three core areas: making robots smarter, broadening the hardware pipeline and scaling production while pushing costs down further.
Still, the post-listing question is larger than simple spending categories. As one of the few A-share names offering direct exposure to humanoid robots, Unitree is an attractive asset for capital markets, but after the listing the main test will be growth execution. The way it deploys fresh capital will directly affect whether its future growth narrative remains credible.
The article cites CCID Media’s 2025 Humanoid Robot Market Research Report, which says China had more than 140 humanoid robot manufacturers in 2025. Shipment volume reached 14,400 units, accounting for 84.7% of the global total, while market size hit 1.55 billion yuan, equal to 53.8% of the world market. It also notes that more than half of Chinese provinces included embodied intelligence and robotics in their 2026 government work reports. That combination of policy support, industrial investment and startup formation points to a rapidly crowding field.
At the same time, the technology bottleneck remains significant. Wei Kai, head of the Artificial Intelligence Institute at the China Academy of Information and Communications Technology, said humanoid robots need a stronger “brain” before they can move into real-world environments at scale. Training data for robotic manipulation remains extremely scarce, which has become one of the biggest constraints on capability development. For Unitree, that means hardware engineering and cost reduction alone will not be enough if model capability and data pipelines lag behind.
Competition is intensifying at home and abroad
Unitree is not competing in an empty market. Domestic names mentioned in the report include EngineAI, AgiBot and Fourier, while international rivals include Tesla’s Optimus and Norway-based 1X with its NEO robot. Fourier already has the GRx series in the market and benefits from an existing rehabilitation robotics business that supports cash flow. Tesla has not mass-produced Optimus Gen 2 yet, but Elon Musk has publicly set a price target near $30,000, putting it in close range of Unitree’s R1 positioning. Meanwhile, 1X NEO has started taking pre-orders and is targeting household scenarios directly.
UBTech is another comparison point frequently raised. Both companies have humanoid products, but their strategic DNA differs. According to the article, UBTech’s 2025 financial report shows its largest revenue contribution came from industrial manufacturing scenarios, suggesting a clearer commercial logic in enterprise deployment. At the same time, UBTech has also opened a new route in emotional companionship. On June 30, it launched the full-size ultra-bionic humanoid “U1 series” globally, with male and female versions, heights ranging from 1.60 meters to 1.85 meters, 88 high-degree-of-freedom joints, and interaction features including dialogue and eye contact. Pricing ranges from 119,800 yuan for the Lite upper-body version to 169,800 yuan for the full-body Pro, and up to 990,000 yuan for the Ultra male version and 880,000 yuan for the Ultra female version. The company said all-channel orders had exceeded 13,000 units.
Three issues will shape Unitree’s path after listing
The article argues that Unitree’s next stage depends on three things. First, can it find a true second category anchor beyond quadrupeds and research-oriented humanoids? The Haneda Airport pilot for G1 is a constructive start, but moving from a trial to signed contracts, and then replicating that model across airports, warehousing, logistics and inspection scenarios, will depend on stability, service infrastructure and delivery capability, not just aggressive pricing.
Second, can Unitree defend its price moat as the market closes in? Today, an R1 at 29,900 yuan and a G1 at 99,000 yuan remain highly disruptive in the sector. But the gap is narrowing as UBTech’s U1 series pushes consumer humanoid pricing lower and Tesla targets a similar broad price band from another angle. Unitree will need to do two things at once: keep reducing costs and turn low price into “low price plus usable performance”.
Third, and most importantly, can it turn embodied AI from a narrative into a delivered capability? The Nvidia-backed H2 Plus, using Jetson Thor and the Isaac GR00T foundation model framework, is intended to support large-scale training of manipulation skills in simulation. Strategically, that direction makes sense. Yet the distance between a lab-ready embodied model and a robot that performs reliably in factories, airports and utility inspections is still greater than many market participants assume. No matter how much hardware is shipped, the ceiling will arrive quickly if perception, reasoning and execution do not improve fast enough.
A likely path: developer tools plus industry landmark deployments
Based on Unitree’s current positioning, the article suggests the company is unlikely to follow UBTech’s consumer companionship route, and also unlikely to replicate Tesla’s “build for your own super factories” logic. A more probable strategy is a mix of developer tooling and vertical benchmark deployments.
Under that framework, Unitree could use lower-priced products such as R1 and G1 to build installed base with developers and universities, allowing external ecosystems to test and refine application scenarios. At the same time, high-performance platforms such as H2 Plus, tightly linked to Nvidia’s stack, could serve as a foundation for embodied AI training. On top of that, the company could rely on showcase projects in airports, factories, power-grid inspection and similar scenarios to enter individual industries one by one.
That route has obvious trade-offs. Volume-led expansion typically weighs on gross margin, while industry projects often involve long sales cycles, customization and slower cash collection. But if the humanoid robot market really expands sharply over the next three to five years, Unitree’s accumulated shipment base, operational data and scenario experience could become strategic assets that are difficult for later entrants to buy or quickly replicate. For now, registration taking effect only means the company has obtained its capital-market ticket. The harder part starts after listing.