August 10, Unitree Goes Public: What the Market Is Really Pricing Is China’s Next Form of Productivity

Originally published on Substack.

On August 10, Unitree Robotics will officially enter the public capital markets.

This will be more than the listing of a robotics company.

It will be a public pricing experiment: How far into the future are investors willing to look—and how much are they willing to pay—for the next upgrade of Chinese manufacturing?

More precisely, the market will not merely be pricing a robot that can run, jump, and perform backflips. It will be pricing China’s ability to combine algorithms, supply chains, data, capital, and national priorities into a new form of productivity.

1. First, Let’s Clarify the RMB 42 Billion Valuation

On July 1, the China Securities Regulatory Commission formally approved the registration of Unitree’s initial public offering. The company plans to issue no fewer than 40.45 million shares, representing at least 10% of its post-IPO share capital, and raise approximately RMB 4.2 billion. A simple calculation—RMB 4.2 billion for a 10% stake—produces the widely cited post-IPO valuation of roughly RMB 42 billion.

But RMB 42 billion is not yet a confirmed market capitalization, much less a final offering price. The ultimate valuation will depend on the book-building process, the offering structure, and investor demand. Treating the fundraising target as the company’s definitive listing valuation is the first and most common mistake in interpreting this IPO.

What matters more are the financial anchors underneath that number.

In 2025, Unitree generated approximately RMB 1.70 billion in revenue, RMB 591 million in adjusted net profit attributable to shareholders, and a gross margin of about 60.44%. Humanoid robots already accounted for 51.78% of revenue. Compared with a field of robotics companies that are still losing money, Unitree’s scarcest asset is not its story. It is the fact that the company has proved it can sell robots—and make money doing so.

The risks, however, are equally visible in the financial statements. In the first quarter of 2026, revenue grew 68.49% year over year, while adjusted net profit fell 52.55%. For the first half, the company expects revenue growth of 35.62% to 45.41%, but a decline in adjusted net profit of 6.43% to 21.97%. Growth remains rapid, but profits are being traded for greater investment in R&D, sales, and organizational expansion.

This means the central question after Unitree’s listing will not be whether robotics has a future. It will be whether today’s earnings can justify the price investors are willing to pay for that future.

Private markets can price the future. Public markets also price disappointment.

2. Unitree Has Four Layers of Value—and Only One Comes From Selling Robots

From an entrepreneur’s perspective, I would divide Unitree’s value into four layers.

The first is product value. Unitree has crossed the divide between laboratory prototype and mass-market product. Its vertically integrated development of motor drives, joint modules, mechanical structures, and motion-control systems enables faster iteration at lower cost. This is the most tangible layer of value and the engineering foundation of the company’s high margins.

The second is platform value. Humanoid and quadruped robots, core components, control systems, and embodied-AI models could form a developer and data flywheel. More robots in the field generate more real-world data. More data improves the models. Better models make it easier for robots to enter new use cases. The problem is that this flywheel remains at an early stage. Unitree itself has disclosed that its general-purpose embodied-AI models have not yet been deployed at scale in its products.

The third is industrial value. A single robot creates demand for motors, reducers, sensors, chips, batteries, precision manufacturing, industrial software, and data services. For local governments and strategic investors, investing in Unitree is not merely a bet on one company. It is a competition to build a new industrial cluster—one that could reproduce some of the supply-chain effects created by electric vehicles.

The fourth is national value. Robotics connects artificial intelligence, advanced manufacturing, labor-force demographics, and technological security. China needs more than a hardware brand that can sell overseas. It needs an autonomous labor-technology stack that can continue to evolve amid restrictions on advanced chips, an aging population, and intensifying global technology competition.

What the state needs is not a robot that can perform backflips, but an industrial chain that can turn algorithms into productivity.

3. Why Unitree Matters So Much to the Chinese Government

In 2023, China’s Ministry of Industry and Information Technology issued its Guiding Opinions on the Innovation and Development of Humanoid Robots, calling for the creation of two or three globally influential ecosystem companies by 2025. In 2025, embodied intelligence appeared in the government work report for the first time. In 2026, the Ministry and the State-owned Assets Supervision and Administration Commission launched real-world training initiatives for humanoid robots and embodied AI, targeting more than 100 high-value use cases and the capacity for deployment at the scale of tens of thousands of units.

Taken together, the policy sequence is clear: support technological breakthroughs, identify national champions, and then use state-owned enterprises, local governments, and industrial environments to help the technology cross the difficult early stages of commercialization.

Unitree’s importance therefore extends well beyond the company itself.

First, it is a showcase for industrial upgrading. Over the past decade, China demonstrated that it could combine policy, infrastructure, supply chains, and market scale into a global competitive advantage in electric vehicles. Robotics is a similar systems-level undertaking, and Unitree is one of the few companies that has already demonstrated cost control, mass-production capability, and global sales.

Second, it is a test case for capital-market reform. Unitree’s IPO moved quickly—from acceptance of its application on March 20, to approval by the listing committee on June 1, and then registration approval in July. For policymakers, the arrival of a profitable, globally visible hard-technology company on the STAR Market helps answer a longstanding question: Can China’s domestic capital markets provide an exit channel for original technology, rather than merely liquidity for mature assets?

Finally, Unitree is a symbol of geopolitical competition. Robots can work in factories, warehouses, and elder-care facilities, but they can also perform inspection, emergency-response, and other high-risk tasks. As their capabilities improve, concerns over data security, export controls, and dual-use technology will become harder to avoid. National-champion status will bring domestic resources, but it may also create a geopolitical discount in overseas markets.

This is the most complex tension in Unitree’s future: The more successful it becomes, the less likely the world is to treat it as an ordinary consumer-electronics company.

4. Why Embodied-AI Companies Are So Easy to Finance in China

According to data compiled by Yicai, China’s embodied-AI sector raised approximately RMB 93.5 billion in the first half of 2026, roughly five times the amount raised a year earlier. Different institutions define “embodied AI” differently, but the acceleration of capital into the sector is unmistakable.

The explanation is not simply that investors believe in robots.

More importantly, embodied AI simultaneously satisfies the needs of four different kinds of capital.

For government funds, it promises industrial clusters, jobs, tax revenue, fixed-asset investment, and technological autonomy. For internet companies, it offers a path for large models to leave the screen and enter the physical world. For automakers and manufacturers, it represents both an extension of their supply chains and a potential workforce for the factories of the future. For financial investors, it is a rare sector with policy support, a vast potential market, and the beginnings of a credible IPO exit channel.

Unitree’s listing will also have a highly practical effect: It will establish a public valuation anchor for the entire industry.

The greatest problem in private markets is not necessarily a high price, but the absence of a comparable price. Once Unitree begins trading, companies such as AgiBot, Galbot, Leju Robotics, DEEP Robotics, and upstream component suppliers will all have a real-time benchmark. Early investors will be able to mark their assets to market. Local-government funds will be able to demonstrate returns on industrial investment. New ventures will find it easier to raise their next rounds.

This is why Unitree matters to Meituan, HongShan, and Matrix Partners China—and why Tencent, Alibaba, Ant Group, China Mobile, and Geely were willing to invest before the IPO. They were not buying only financial returns. They were also buying supply-chain synergies, access points for models and cloud services, deployment opportunities, industry intelligence, and a seat in the next-generation robotics ecosystem.

From a political-economy perspective, the embodied-AI financing boom is a migration from one capital container to another. Real estate is deleveraging. Consumer internet has entered a zero-sum phase. Pure software faces both business-model pressure and commoditization. Embodied AI, by contrast, can absorb land, factories, equipment, computing power, engineers, and industrial funds at the same time.

What makes embodied AI so attractive to capital is not that it resembles AI, but that it can turn AI back into factories, orders, and assets.

5. What Could Happen to Unitree’s Valuation After the Listing?

Using Unitree’s 2025 revenue of RMB 1.70 billion and adjusted net profit of RMB 591 million, a valuation of RMB 42 billion implies approximately 24.7 times sales and 71 times adjusted earnings. Using reported net profit attributable to shareholders of RMB 278 million, the price-to-earnings multiple exceeds 150. The gap largely reflects a one-time share-based compensation expense. It also raises an important question for investors: Is equity compensation truly a one-time item, or is it an enduring cost of attracting talent at a technology company?

Rather than offer a deceptively precise price target, I prefer to think in three scenarios.

ScenarioMarket-Cap RangeHow the Market Views UnitreeSignals to WatchValuation compressionRMB 28–35 billionA high-growth hardware company that deserves a discount for declining profits, rising R&D spending, and intensifying competitionWeak repeat orders from industrial customers; continued profit erosion; tighter overseas restrictions; rising receivables and inventoryBase-case pricingRMB 42–55 billionA scarce, profitable category leader and the first publicly traded humanoid-robot pure play in the A-share marketSustained revenue growth; stable gross margins; delivery and mass-production targets met; model development progressing as plannedSpeculative exuberanceRMB 65–80 billionIPO scarcity, policy narratives, and robotics enthusiasm combine to push the valuation sharply higherHeavy oversubscription, tight first-day liquidity, and a broader sector rally—but sustained performance would still require real customer orders

In the short term, Unitree has the ingredients for a strong opening and extreme volatility. During the first half of 2026, newly listed A-shares rose by an average of roughly 280% on their first trading day, with hard-technology companies on the STAR Market attracting particular enthusiasm. But that figure measures market sentiment, not Unitree’s intrinsic value, and it cannot be mechanically extrapolated into a forecast for the stock.

Over the medium term, RMB 42 billion may be only the beginning of the valuation debate. Every quarter, investors will ask three questions: Are Unitree’s robots performing real work, or are they still concentrated in research, education, and demonstration settings? Can the company preserve its high humanoid-robot margins in a price war? And can the billions of renminbi invested in the robot “brain” evolve from an expense into software, data, and ecosystem revenue?

Over the long term, Unitree’s ability to sustain a valuation above RMB 60 billion will not depend on how many spectacular moves its robots can perform. It will depend on whether those robots can work continuously, generate repeat orders, reduce labor requirements, and create a closed data loop.

Valuation can build a bridge between policy and profit, but the stock price must ultimately return to cash flow.

6. What Entrepreneurs Should Really Learn

The most important lesson of Unitree’s IPO is not to “choose a hot sector.” It is to build the capacity to survive in a long-term technological direction—and then wait for the macro narrative to catch up.

Unitree was founded in 2016. Embodied AI did not become a policy and investment buzzword until years later. The company first used quadruped robots to build revenue, supply-chain capabilities, and brand recognition before moving into humanoids. It strengthened the body and the “cerebellum” before investing heavily in the “brain.” It sold products to global customers before asking capital markets to recognize its platform value.

This is a difficult path to replicate, but an important one to understand: Use products that can be sold today to finance technologies for the distant future. Use engineering capability to give a capital-markets narrative a foundation in reality.

Going public, however, changes the rules of the game. A startup can tolerate changes in direction; a public company must explain every quarter. A founder can believe in a ten-year vision; public investors will ask about this quarter’s earnings. Governments can open use cases; customers will still calculate their return on investment.

Unitree’s real test is not the listing itself. It is whether the company can preserve the speed of a startup while learning the discipline of a public company.

If it succeeds, China will gain more than a robotics champion. It will gain a growth flywheel capable of recirculating capital, manufacturing, data, and AI.

If it fails, the IPO will still have value. It will force the entire industry, for the first time, to test the most expensive technology narrative of recent years against public financial statements, real orders, and sustained cash flow.

Unitree’s IPO is not the end of its fundraising journey. It is the beginning of China’s embodied-AI sector moving from a “narrative asset” to an “operating asset.”

August 10 will mark that beginning. But for entrepreneurs and investors, a more important question has already emerged:

When policy, capital, and technology are all standing behind you, can you convert the valuation your era has given you into value that customers will continue to pay for?