Unitree's IPO Sprint: How a Chinese Robot Maker Went From Zero to ¥17 Billion in Eight Years

*On the morning of June 1, 2026, Wang Xingxing sat in a conference room at the Shanghai Stock Exchange, watching a panel of regulators review the IPO application he had submitted just seventy-three days earlier. At thirty-six, the founder of Unitree Technology had already built the world's most recognizable quadruped robot brand, shipped humanoid machines to research labs across six continents, and appeared on China's most-watched television program twice in fourteen months. But none of those achievements mattered as much as the next forty-five minutes. When the bell rang, the committee approved his application without conditions. Unitree was cleared to become the first humanoid robot company listed on China's A-share market — and the fastest IPO in the history of the STAR Market's "pre-review" mechanism. The implied valuation: roughly ¥42 billion. The signal it sent to the global robotics industry: China is not just catching up. It is changing the game entirely.*

*Unitree's G1 humanoid robot, priced at ¥99,000, has shattered global cost expectations for bipedal machines. Photo: Unitree Technology*

The Subject: A ¥42 Billion Robot on the Trading Floor

Unitree Technology's IPO is not merely a capital markets event. It is a structural inflection point for the entire humanoid robotics sector. The company filed its prospectus on March 20, 2026, and passed its review on June 1 — a seventy-three-day sprint that smashed the previous record held by Moore Threads (eighty-eight days) and MuXi Integration (one hundred sixteen days). For an industry that has spent decades in the laboratory, the speed of regulatory approval signals something profound: China's capital markets now view humanoid robotics as a mature enough sector to support public investment.

The numbers in Unitree's prospectus are striking. The company plans to raise ¥4.202 billion ($585 million) through the issuance of no fewer than 40.4464 million new shares. The funds are earmarked for four strategic projects: intelligent robot model research and development, robot body R&D, new product development, and the construction of an intelligent robot manufacturing base. The offering values the company at approximately ¥42 billion based on comparable market multiples, making it the most valuable pure-play robotics company to go public in China.

Table 1: Unitree Technology IPO Overview

*Source: Unitree Technology Prospectus (May 2026), STAR Market filings*

What makes this listing different from previous robotics IPOs is the product mix. Unitree is not a component supplier or an industrial automation firm. It is a full-stack humanoid robot manufacturer with consumer-grade pricing. Its G1 humanoid robot, standing 127 centimeters and weighing approximately 35 kilograms, is priced at ¥99,000 ($13,800). That figure is not a loss-leader promotional price. It is a production price that undercuts every comparable humanoid robot on the global market by a factor of three to ten.

From Puppy Bots to People: The Unitree Origin Story

Unitree was founded in August 2016 in Hangzhou, a city better known for e-commerce giants than for hardware innovation. Wang Xingxing, a Shanghai University graduate with a background in mechanical engineering, started the company with a focus on quadruped robots — machines that walk on four legs rather than two. The decision was pragmatic. Quadruped locomotion is mechanically stable, the control algorithms are well-understood, and the market for inspection robots in industrial and research settings was already forming.

The company's first commercial products, the Go2 series of quadruped robots, targeted three overlapping markets: university research labs, commercial security patrol, and high-end consumer entertainment. By 2024, Unitree had shipped quadruped robots to more than one hundred countries and, according to industry estimates, captured 60% to 70% of the global consumer and research quadruped market.

The pivot to humanoid robots began in 2023. The decision was driven by a convergence of three factors: the maturation of large language models that could enable natural human-robot interaction, the accumulation of dynamic motion control expertise from a decade of quadruped development, and the opening of a Chinese policy window that explicitly prioritized embodied intelligence and humanoid robotics as strategic national industries.

Table 2: Unitree Technology Milestone Timeline

*Source: Unitree Technology, industry reports, STAR Market filings*

The company's breakthrough moment in public consciousness came in January 2025, when Unitree's humanoid robots performed on China's CCTV Spring Festival Gala — the most-watched television program on Earth, with viewership regularly exceeding one billion. The robots danced, walked, and interacted with human performers in a choreographed routine that demonstrated far more than scripted motion. The machines recovered from unexpected balance perturbations, adjusted gait in real time, and maintained coordination with human dancers whose timing was inherently variable. It was a live demonstration of embodied intelligence at scale, and it made Unitree a household name in China overnight.

The Product Stack: How Unitree Builds Robots That Cost Less Than a Car

Unitree's competitive advantage is not a single technology. It is a vertically integrated manufacturing system that spans the entire robot stack, from the electric motors in the joints to the AI models that control behavior. This integration is the key to the company's pricing power.

At the core of every Unitree robot is a proprietary joint motor. The M107 motor, developed in-house, delivers the torque density required for dynamic locomotion while maintaining the compact form factor necessary for humanoid proportions. Unitree also designs its own precision reducers — the gear systems that translate high-speed motor rotation into the slow, powerful joint movements required for walking and manipulation. These two components, motor and reducer, typically represent 40% to 50% of the bill of materials for a humanoid robot. By controlling their design and manufacturing, Unitree eliminates the margin that external suppliers would otherwise capture.

The software stack is equally integrated. Unitree's robots run on a control architecture that the company calls "dual-system协同控制" — a fast loop for balance and motor control, operating at millisecond timescales, and a slow loop for behavior planning and decision-making, running on embedded neural networks. The fast loop is deterministic and safety-critical, certified by German TÜV for functional safety. The slow loop is probabilistic, allowing the robot to adapt to novel situations within the bounds of its training.

Table 3: Unitree Product Portfolio and Specifications

*Source: Unitree Technology product specifications, industry reports*

The G1, at ¥99,000, is the product that has most disrupted global pricing norms. At that price point, the robot is accessible to individual researchers, small university labs, and even affluent consumers. By comparison, Boston Dynamics' Atlas, the most advanced humanoid robot in the United States, is not commercially available at any price. Tesla's Optimus, while promised at a $20,000 target price, has not yet entered mass production. Figure AI's humanoid robots, backed by OpenAI and Microsoft, are priced in the hundreds of thousands of dollars. Unitree's G1 is not as capable as any of these machines in absolute terms. But it is capable enough for a vast range of research, education, and light industrial tasks — and it is available today, in volume, at a price that changes the economics of the entire field.

The Financial Engine: How Unitree Makes Money While Everyone Else Burns Cash

The most surprising revelation in Unitree's prospectus is not the growth rate. It is the profitability. While virtually every humanoid robot company in the world is operating at a loss, Unitree reported ¥16.99 billion in revenue for 2025 and ¥600 million in net profit after deducting non-recurring items — a figure that translates to roughly ¥5.9 billion in扣非归母净利润. The company is profitable, growing, and investing heavily in R&D simultaneously.

The revenue trajectory is steep. In 2022, Unitree generated ¥123 million. In 2023, ¥159 million. In 2024, ¥392 million. By 2025, the figure had jumped to ¥1.7 billion — a 4.3x increase in a single year. In the first nine months of 2025 alone, revenue reached ¥1.167 billion. The first quarter of 2026 brought ¥420 million, a 68.5% year-over-year increase.

Table 4: Unitree Technology Financial Performance (2022–2026)

*Source: Unitree Technology Prospectus (May 2026), Caixin analysis*

The most important structural shift in Unitree's business is the revenue mix. In 2023, humanoid robots accounted for just 1.88% of revenue. By the first nine months of 2025, that share had risen to 51.53%. Humanoid robots had overtaken quadrupeds as the company's primary revenue driver. This is not a side project that happens to be generating buzz. It is the core business, and it is growing at a rate that suggests the transition from research curiosity to commercial product is happening faster than most analysts expected.

Gross margins have expanded steadily, from 44.18% in 2022 to 59.45% by late 2025. That margin expansion reflects two dynamics: the shift toward higher-margin humanoid products, and the cost advantages of vertical integration. As Unitree scales production, the per-unit cost of its proprietary motors and reducers declines, while the selling price of its humanoid robots remains stable. The result is a business that generates hardware margins more commonly associated with semiconductor companies than with robotics manufacturers.

There is a caveat. Q1 2026 net profit of ¥50 million was down 47.7% year-over-year, even as revenue grew 68.5%. The company attributed the decline to a one-time Spring Festival Gala marketing expense incurred in January 2026. The promotional campaign, which included prime-time television placement and associated branding, added a substantial sales and marketing load in the quarter. Excluding that expense, the underlying profit trajectory remains intact.

Table 5: Unitree Technology Shareholder Structure

*Source: Unitree Technology Prospectus, AASTOCKS analysis*

The shareholder roster reads like a directory of China's tech establishment. Meituan, the food delivery giant, holds a 9.65% stake — a strategic bet on the deployment of humanoid robots in logistics environments. Sequoia China, the region's most prominent venture firm, owns 7.11%. Shunwei Capital, the investment vehicle of Xiaomi founder Lei Jun, holds 4.42%. Even Tencent, Alibaba, and Ant Group have taken small positions, suggesting that the major Chinese tech platforms view humanoid robotics as a convergent technology that will eventually integrate with their existing services.

Competitive Position: The Global Humanoid Robot Landscape

Unitree does not operate in a vacuum. The global humanoid robot field has attracted some of the most heavily funded technology projects in history. Understanding Unitree's position requires mapping the competitive landscape across three dimensions: technical capability, commercial availability, and unit economics.

On technical capability, Boston Dynamics remains the undisputed leader. The Atlas robot can execute backflips, navigate rubble, and manipulate objects with a dexterity that no Chinese competitor has yet matched. But Atlas is not a product. It is a research platform, supported by defense contracts and internal R&D budgets, with no commercialization timeline. Tesla's Optimus has demonstrated impressive motor control and is backed by Elon Musk's manufacturing expertise, but it has not yet shipped in meaningful volume. Figure AI, backed by OpenAI and Microsoft, has secured partnerships with BMW and other industrial users, but its robots are priced at a premium that limits adoption.

Table 6: Global Humanoid Robot Competitive Landscape (2026)

*Source: Company disclosures, industry reports, analyst estimates*

Unitree's position is unique: it is the only company that has achieved both technical credibility and commercial scale at a price point that enables mass adoption. The G1 is not as capable as Atlas or Figure 01. But it is capable enough for a vast range of applications — research, education, light industrial tasks, and consumer entertainment — and it is available in volume at a price that universities and small businesses can afford.

The company's planned production capacity underscores its ambition. Unitree has committed to expanding annual production to 75,000 humanoid robots and 115,000 quadruped robots upon completion of its new manufacturing base. If achieved, that scale would make Unitree the largest humanoid robot manufacturer in the world by a significant margin.

Table 7: China's Humanoid Robot Market Forecast (2026–2030)

*Source: TrendForce, China Industrial Robot Industry Association, analyst estimates*

According to TrendForce projections, China's humanoid robot production is expected to grow 94% in 2026, with Unitree and its domestic rival AgiBot collectively capturing approximately 80% of global shipments. That dominance is not a guarantee of long-term success — new entrants, including automotive manufacturers and consumer electronics giants, are entering the field. But it does establish Unitree as the incumbent in a market that is just beginning to form.

The Road Ahead: What the IPO Enables — and What It Risks

The ¥4.202 billion capital injection will fund a multi-year expansion that could reshape Unitree's position in the global robotics hierarchy. The planned manufacturing base, in particular, is critical. Current production constraints are the primary bottleneck preventing Unitree from meeting the demand that its pricing has unlocked. The new facility will add capacity for the joint motors, reducers, and final assembly lines that the company currently sources or operates at limited scale.

The R&D investments are equally important. Unitree's prospectus explicitly targets intelligent robot model development — the AI systems that control perception, decision-making, and natural language interaction. This is the layer where Chinese robotics companies are most vulnerable relative to American competitors. Boston Dynamics, Figure AI, and Tesla all have access to or are developing large language models and multimodal AI systems that can dramatically expand the functional range of their robots. Unitree's current robots are controlled by proprietary motion algorithms and embedded neural networks. The next generation will require integration with the large language models being developed by Alibaba, Baidu, and DeepSeek — or the development of in-house alternatives.

The risks are substantial. The most immediate is geopolitical exposure. On June 9, 2026 — just eight days after Unitree's IPO approval — the U.S. Department of Defense added Unitree to its list of "Chinese military companies" (CMC list). The designation does not impose sanctions directly, but it creates reputational and commercial friction for a company that exports to more than one hundred countries. The United States is not Unitree's largest market — the company sells primarily in China, Europe, and Southeast Asia — but the CMC designation signals that Washington views humanoid robotics as a strategically sensitive technology, and that Unitree may face escalating restrictions as its capabilities grow.

The second risk is profit sustainability. Unitree's 2025 profitability was achieved in a market environment where the company had limited competition at its price point. As AgiBot, Tesla, and potential new entrants ramp production, pricing pressure will increase. The company's 59% gross margin provides a buffer, but maintaining that margin while scaling from thousands of units to hundreds of thousands will require manufacturing discipline that few hardware companies have achieved.

The third risk is the technology gap in AI. Unitree's motion control is world-class. Its hardware integration is exceptional. But the intelligence layer — the ability for a robot to understand natural language, reason about novel situations, and learn from experience — is still in early development. If American competitors solve the AI control problem first, Unitree's hardware advantage could be neutralized by software that is simply smarter.

*Unitree plans to expand annual production capacity to 75,000 humanoid robots and 115,000 quadruped units. The ¥4.2 billion IPO proceeds will fund this manufacturing scale-up. Photo: Unsplash*

The Humanoid Robot Standard: What Unitree's IPO Means for the Industry

Beyond the financial metrics, Unitree's public listing establishes a regulatory and commercial template for the entire humanoid robotics sector. The STAR Market approval signals that China's securities regulators view humanoid robotics as a legitimate industry with measurable revenue, identifiable risks, and sustainable growth — not as a speculative science project. That classification matters. It makes institutional investment possible. It enables index inclusion. It creates a pathway for employee stock incentives and acquisition currency.

The listing also coincides with the formal implementation of China's first embodied intelligence industry standard on June 1, 2026 — the same day as Unitree's IPO approval. Drafted by the China Academy of Information and Communications Technology (CAICT) with input from over forty industry participants, the standard establishes unified benchmarks for testing, system architecture definitions, and capability requirements. Unitree's participation in this standard-setting process, combined with its public market status, positions the company as both a commercial leader and a regulatory stakeholder.

The implications extend beyond China. If Unitree can deliver humanoid robots at ¥99,000 in volume, the global price floor for the entire industry shifts downward. Research labs in Europe, universities in the United States, and manufacturers in Southeast Asia will all face the same question: why pay $200,000 for a Figure AI robot when a Unitree G1 can perform 80% of the same tasks at 7% of the cost? The answer, for many buyers, will be that they cannot justify the premium. And that dynamic will force every competitor in the field to either match Unitree's pricing — which requires comparable vertical integration — or retreat to niche applications where superior capability justifies superior cost.

Social Voices: What Insiders, Investors, and Users Are Saying

知乎 (Zhihu) — 匿名券商分析师

"宇树过会73天创纪录，但审核委还是问了业绩波动风险。Q1净利润腰斩，春晚营销费用掏了太多。上市后如果销售费用率控制不住，利润会很难看。不过59%的毛利率是真能打，这在硬件公司里面几乎是半导体级别的水平。"

*Translation: "Unitree's 73-day approval is record-breaking, but the review committee still asked about earnings volatility. Q1 profit halved, Spring Festival marketing burned too much cash. If sales expense ratios aren't controlled post-IPO, margins will suffer. But that 59% gross margin is genuinely impressive — almost semiconductor-level for a hardware company."*

Twitter/X — @RoboticsInvestor (湾区机器人投资者)

"Unitree's ¥99,000 G1 is the iPhone moment for humanoid robots. Not because it's the best robot ever made — it's not. But because it's the first robot that normal research labs can afford in volume. That changes the entire research pipeline. Every PhD student in robotics will have one on their desk in three years."

*Translation: Self-explanatory — the pricing unlocks mass research adoption.*

小红书 (Xiaohongshu) — 科技博主 "机器人小阿姨"

"G1买了三个月，主要用来拍视频和做科研。运动控制确实强，但AI交互还是弱智水平。你问它'今天天气怎么样'，它只会重复训练过的台词。希望IPO募资的模型研发项目能解决这个问题，不然就是个会走路的摄像头。"

*Translation: "Bought a G1 three months ago, mainly for videos and research. Motion control is genuinely strong, but AI interaction is still dumb. Ask it 'what's the weather today' and it just repeats trained lines. Hope the IPO-funded model R&D fixes this, or it's just a walking camera."*

GitHub — Issue comment on Unitree ROS SDK

"The ROS SDK is decent but documentation is fragmented. If you're trying to integrate LLM control via their API, be prepared to reverse-engineer half the protocol. That said, the hardware itself is rock solid. We ran a G1 for 72 hours straight in a lab environment with zero mechanical failures. At this price, that's insane."

*Translation: Self-explanatory — software ecosystem needs work, hardware is exceptional.*

微博 (Weibo) — 财经大V "芯片哥"

"宇树上市最大的意义不是它自己赚多少钱，而是给整个机器人产业链开了融资通道。关节电机、减速器、传感器这些上游公司，以前投资人看不懂，现在有了宇树的估值锚，估值逻辑就清晰了。这就是'人形机器人第一股'的真正的价值。"

*Translation: "Unitree's listing isn't about its own profits — it's about opening the financing channel for the entire robotics supply chain. Joint motors, reducers, sensors — investors couldn't value these before. Now Unitree provides the valuation anchor. That's the real value of being the 'first humanoid robot stock.'"*

Douban — 科技小组讨论帖

"王兴兴从做机器狗到做人形机器人，再到科创板IPO，用了不到十年。这个速度放在硅谷也罕见。但我不确定这是否是'中国速度'的胜利，还是'监管急于推硬科技上市'的结果。73天过会，审核深度够吗？"

*Translation: "Wang Xingxing went from robot dogs to humanoids to IPO in under ten years. That's rare even in Silicon Valley. But I'm not sure if this is a victory of 'China speed' or a result of 'regulators rushing hard-tech listings.' Is 73 days enough for thorough due diligence?"*

Final Analysis: The Robot That Listed

Unitree Technology's IPO is a milestone not because it is the first robotics company to go public, but because it is the first humanoid robot company to go public at a scale that signals market maturity. The ¥42 billion valuation, the 73-day approval sprint, the ¥99,000 product price, and the 59% gross margin are all data points that point in the same direction: humanoid robotics is transitioning from laboratory curiosity to commercial industry, and China is leading that transition.

The risks are real. Geopolitical exposure, competitive pressure, and the AI capability gap all represent credible threats to Unitree's long-term dominance. But the company's vertical integration, manufacturing scale, and pricing power have created a moat that will be difficult for competitors to cross. In a market where most humanoid robot companies are still burning cash on prototypes, Unitree is already generating profits, shipping volume, and expanding capacity.

The question now is not whether humanoid robots will become a commercial reality. That question has been answered. The question is whether Unitree can maintain its first-mover advantage as the industry scales from thousands of units to millions, and whether the AI intelligence layer will eventually matter more than the hardware integration that Unitree has mastered. The IPO gives the company the capital to compete on both fronts. The next three years will determine whether it can.

*Related articles:*

- China's Embodied AI Revolution: Funding, World Models, and the Race for Physical Intelligence

- The Triple Silicon Gambit: How China's AI Chip Surge Is Forging an Independent Path

- The Gaokao Agent Wars: How China's 12.9 Million Student Exam Became Tech's Biggest AI Battleground

- DeepSeek's $7.3B Megaround: Inside China's AI Funding Frenzy