Unitree IPO Is Real, but the Gray-Market Trading Story Is Not What It Seems

Unitree Robotics has advanced a real IPO, yet the supposed gray-market trading around it involves a very different and far riskier product.

The Hangzhou robot maker opened subscriptions for its Shanghai STAR Market offering on August 10, 2026. However, Unitree shares had not begun official exchange trading by August 15. Products marketed elsewhere as pre-IPO exposure were therefore not ordinary Unitree stock.

That distinction matters because some trading platforms launched contracts using Unitree’s name before its public-market debut. These instruments can resemble early access to a sought-after listing. In practice, buyers receive a platform-defined derivative whose price can detach from the eventual A-share.

The Unitree IPO itself is no rumor. Regulators approved it, formal subscription procedures began, and the company received the stock code 688836. The speculation begins when an unrelated trading venue presents its own contract as a meaningful preview of Unitree’s market value.

This is the central conflict: a regulated securities offering is being conflated with a leveraged proxy created outside that offering. The first grants shares through established allocation and settlement processes. The second grants contractual exposure under rules written by a trading platform.

Unitree’s business also deserves more scrutiny than its viral robot demonstrations usually receive. Its revenue grew rapidly during 2025, but profitability weakened in early 2026. Commercial deployment remains concentrated in research, education, demonstrations, reception, and limited industrial applications.

The IPO offers investors a rare chance to value a profitable humanoid robot manufacturer. It also asks them to price technical uncertainty, geopolitical exposure, concentrated control, and a crowded Chinese robotics market.

The Unitree IPO Has Cleared the Regulatory Process

Unitree’s public offering is documented, but official documentation does not turn an outside derivative into company stock.

Unitree filed its Shanghai listing application on March 20, 2026. The Shanghai Stock Exchange’s listing committee reviewed and approved the application on June 1. China’s securities regulator approved the registration in early July.

The company then published its formal offering arrangements on July 31. Those documents established a preliminary inquiry on August 5 and subscriptions on August 10.

Unitree planned to issue about 40.45 million new shares. That amount represents 10 percent of its enlarged share capital. Existing shareholders were not selling shares through the offering.

The offer combines strategic placement, institutional allocation, and online public subscription. The lead underwriter organizes those channels under Shanghai Stock Exchange procedures. Successful subscriptions ultimately settle through China’s securities infrastructure.

Those mechanics define the actual Unitree IPO. A legitimate allocation produces registered A-shares after payment, settlement, and listing. It does not produce a token, perpetual future, or private platform balance.

An IPO subscription is also different from exchange trading. Subscription lets eligible investors request shares before listing. It does not create continuous trading in those shares before the exchange opens them.

As of August 15, the reviewed offering documents established the subscription timetable but did not support claims of official pre-debut trading. A formal listing announcement remains the authoritative source for the first trading date.

The distinction can be confusing because “gray market” has several meanings. In some financial centers, brokers facilitate conditional trading before settlement. Those trades still concern the expected securities and operate within specific market conventions.

The Unitree-linked contracts promoted online do not fit that simple description. At least one cryptocurrency venue advertised a pre-IPO futures product offering round-the-clock trading and leverage before Unitree’s listing.

That product sits outside the STAR Market allocation. Unitree did not issue it, and the Shanghai Stock Exchange does not operate it. Its existence therefore says nothing about whether investors can legally trade Unitree A-shares.

The stock code can make the imitation more persuasive. Unitree’s official code is 688836, while an outside contract can use a recognizable label without conferring ownership. A familiar name does not establish a legal claim on shares.

Investors should trace any alleged position through four questions. Who issued the instrument? What asset settles it? Which regulator oversees the venue? What happens if the official listing is delayed?

For an actual allocation, offering documents answer those questions. For a synthetic contract, the platform’s own terms provide the answers. Those terms can differ significantly from securities-market protections.

The safest conclusion is narrow. Unitree has completed major IPO milestones, and subscriptions have occurred. That does not verify every product advertised as Unitree exposure before exchange trading begins.

Why Unitree Gray-Market Trading Is a Misleading Label

The controversial contract is a wager on a future reference price, not evidence that Unitree shares already trade in secret.

A futures contract creates an obligation or cash-settled exposure based on a specified reference value. It does not necessarily deliver the underlying asset. A perpetual future may lack a fixed expiration date altogether.

MEXC promoted UNITREE pre-IPO futures before the Shanghai debut. Its product announcement advertised early exposure, continuous trading, and leverage reaching 25 times.

Those features reveal the product’s actual character. Unitree’s A-share does not trade continuously, and STAR Market subscriptions do not offer that leverage. The contract follows the venue’s trading engine and settlement policy.

Calling this activity a gray market gives it an aura of privileged price discovery. That framing implies informed investors are quietly determining Unitree’s value before ordinary buyers arrive.

The inference is too strong. A small derivative market can reflect enthusiasm among cryptocurrency traders rather than demand from qualified institutional investors. Its participants can differ from the institutions involved in Shanghai’s book building.

Liquidity matters as much as branding. A contract with a thin order book can move sharply after modest buying or selling. Its quoted price might therefore reveal more about local positioning than Unitree’s expected valuation.

Leverage amplifies the problem. At 25 times leverage, an adverse move of roughly four percent can consume the trader’s posted margin before fees and platform rules. Liquidation can occur without any change in Unitree’s operations.

The contract also introduces basis risk. Basis risk is the difference between a derivative’s price and the value it supposedly tracks. That gap can remain wide when the underlying security is not yet available for arbitrage.

Ordinary arbitrage helps align a derivative with its underlying asset. Traders buy the cheaper instrument and sell the expensive one. Before Unitree shares trade, that correcting mechanism cannot operate normally.

Settlement language becomes critical. The platform must decide which official price, date, or conversion formula governs the contract after listing. Delays, symbol changes, or trading restrictions can complicate that process.

Counterparty and venue risks remain separate from Unitree’s business. A trader can judge Unitree correctly yet lose because of liquidation rules, outages, contract amendments, or disputes over the reference price.

There is also a jurisdiction problem. Unitree is pursuing a mainland Chinese A-share listing. Access to STAR Market securities depends on investor eligibility, brokerage arrangements, and cross-border market rules.

A crypto derivative does not remove those restrictions. It replaces direct ownership with a claim against a platform. That substitution can expand accessibility while reducing legal clarity.

The product might still serve speculative demand. Traders routinely use event contracts to express views before an asset becomes available. The problem begins when marketing or social posts blur speculation with ownership.

No observed contract price should be called Unitree’s official share price before the stock starts exchange trading. It is only the market price of that particular contract.

That makes the “dark trading” story partly true but materially misleading. Something bearing the Unitree name was trading. Unitree’s registered A-shares were not secretly changing hands through that venue.

The Unitree IPO Tests Revenue Quality, Not Robot Choreography

Unitree has delivered exceptional growth, but public investors must decide how much of that growth can survive outside demonstrations and research laboratories.

Unitree reported nearly 1.7 billion yuan in 2025 revenue, up from 393 million yuan during 2024. Its operating income increased 335 percent, according to the company’s prospectus.

Humanoid robots became the largest business segment during the first nine months of 2025. Their share of core revenue reached 51.5 percent, compared with 27.6 percent in 2024.

The company says it shipped more than 5,500 humanoid robots during 2025. Its filing attributed 32.4 percent of the global humanoid market to those shipments.

These numbers establish Unitree as more than a robotics demonstration company. It has built repeatable hardware, production systems, distribution channels, and a recognized international brand.

However, unit shipments do not automatically prove broad commercial usefulness. A robot sold to a university can generate revenue without demonstrating that the machine performs profitable work inside an uncontrolled factory.

Unitree’s filing says research and education account for much of its demand. Reported industry applications include reception, guided tours, manufacturing, and inspection.

Reception and tour-guide uses represented roughly 50 to 70 percent of humanoid industry-application revenue, according to the IPO filing analysis.

These deployments provide genuine customer experience, but they require less autonomy than many proposed humanoid applications. Greeting visitors differs greatly from manipulating varied objects across an eight-hour industrial shift.

Embodied AI refers to artificial intelligence acting through a physical machine. The robot must perceive its surroundings, plan movements, and complete tasks without unsafe or expensive failures.

A stage routine can be planned, rehearsed, and tightly controlled. A warehouse, hotel, or factory presents moving people, irregular objects, uncertain lighting, and changing layouts.

Reliability therefore becomes more valuable than a dramatic demonstration. Enterprise buyers care about uptime, maintenance intervals, intervention rates, safety records, and total deployment costs.

Unitree’s lower-priced G1 helped broaden demand, but the product mix reduced gross margin. That tradeoff illustrates a recurring hardware problem: lower prices increase adoption while placing more pressure on component and service economics.

The company’s planned fundraising targets robot AI models, robot-body research, new products, and manufacturing capacity. It is seeking about 4.2 billion yuan for these projects.

That investment could strengthen the integration between hardware and software. It could also raise fixed costs before a stable market for general-purpose humanoid work has developed.

The real Unitree IPO debate is therefore not whether the robots can dance, run, or perform martial arts. Those capabilities show sophisticated motion control and mechanical engineering.

The harder question concerns repeatable economic output. How many customers reorder after a pilot? How much supervision does each robot need? How frequently do components require replacement?

Developers and enterprise buyers should track those operating measurements. Demonstration videos establish possibility, while sustained deployments establish value.

This distinction resembles the difference between a model benchmark and a production AI workflow. A benchmark isolates performance. Real deployment adds data quality, integration, monitoring, exceptions, and user behavior.

Teams evaluating robotics claims need a reliable evidence trail across filings, demos, test results, and customer reports. A searchable AI knowledge base can keep those sources connected without treating each announcement as an isolated event.

Fast Growth Has Not Removed the Earnings Risk

Unitree enters the market with real revenue and profit, yet its early 2026 results show how quickly competition can pressure that story.

Unitree reported 1.708 billion yuan in 2025 revenue. Adjusted profit reached about 590 million yuan, according to the prospectus and exchange-related disclosures.

That profitability separates Unitree from several robotics peers. UBTech generated about 2 billion yuan in 2025 revenue but recorded a loss near 700 million yuan.

The comparison makes Unitree look unusually efficient. It also raises expectations that might become difficult to meet as spending, competition, and international restrictions increase.

During the first quarter of 2026, Unitree’s revenue rose 68 percent to 423 million yuan. Adjusted profit fell 53 percent to approximately 40 million yuan.

The company attributed the decline to higher research spending, marketing expenses, and product price reductions. Those pressures are not temporary accounting curiosities. They reflect the economics of a young hardware category.

More manufacturers are entering humanoid robotics. Rivals compete for engineering talent, components, training data, customers, and public attention.

China already has multiple capital-market paths forming around the sector. UBTech trades in Hong Kong. Leju Robotics and DEEP Robotics have pursued domestic listings, while AgiBot selected another capitalization route.

This competition can lower hardware prices before software capabilities become sufficiently differentiated. Buyers benefit initially, but manufacturers must absorb lower margins or reduce spending.

Unitree’s brand gives it an advantage. Its robots appear frequently in public demonstrations, sporting events, laboratories, and entertainment programs. That visibility lowers customer-acquisition friction.

Brand recognition does not eliminate switching. Enterprise purchasers will compare task completion, integration, safety, maintenance, and support. A memorable performance does not guarantee the best industrial economics.

Reuters Breakingviews argued that existing humanoids still struggle with dexterity and intelligence during basic uncontrolled tasks. Its robotics risk analysis highlighted pouring water and connecting cables as continuing challenges.

Unitree itself acknowledges limited commercial applications for humanoids and quadruped robots. That disclosure matters more than promotional claims about a general-purpose robotic workforce.

The company can continue growing while the investment thesis weakens. Revenue might rise through lower-priced units, while margins decline because each sale requires more support or carries less hardware profit.

Investors should therefore separate three measures. Shipment growth shows demand. Gross margin reflects hardware economics. Repeat deployments indicate whether customers receive durable operational value.

The IPO’s high visibility can distort that evaluation. Scarcity, national technology ambitions, and retail demand can support valuation before deployment data catches up.

Unitree also uses a dual-class structure. Founder Wang Xingxing and a controlled employee platform are expected to hold 31.29 percent of shares after issuance.

Their voting power is expected to remain about 65.31 percent. The structure supports long-term founder control, but public shareholders gain limited influence over strategy and governance.

That arrangement is neither automatically good nor bad. It simply concentrates responsibility. If Unitree makes an expensive technical bet, outside shareholders have fewer tools to change direction.

A synthetic pre-IPO contract adds another layer of separation. Its holder lacks both ordinary shareholder protections and direct influence. The trader receives price exposure without the underlying governance rights.

Overseas Exposure Turns Policy Into a Product Risk

Unitree’s international business is meaningful enough that trade restrictions can affect product roadmaps, not merely investor sentiment.

Overseas markets generated more than 40 percent of Unitree’s revenue during each reporting period disclosed in its offering materials. The United States represented a significant portion.

Unitree reported that US sales contributed 18.39 percent, 19.54 percent, and 13.30 percent across the three relevant periods. That exposure gives American policy direct financial importance.

In late July, the Federal Communications Commission added foreign-made advanced robots to its Covered List. New affected devices face restrictions on equipment authorization needed for US sales.

Existing authorized models can remain available under the current framework. Future Unitree products might require an exemption or conditional approval, according to the company’s warning.

The distinction creates an unusual product-cycle risk. Unitree can continue selling certain existing devices while losing access for a future model.

That outcome could slow upgrades, fragment regional product lines, or leave US customers using older hardware. It could also shift development resources toward markets with fewer authorization barriers.

Unitree warned that further tightening might affect existing approvals. Reuters summarized these risks in its US exposure report.

Trade restrictions can also affect components, cloud services, software updates, and research partnerships. Robotics products combine sensors, processors, actuators, networking, and AI systems from complicated supply chains.

A restriction in one layer can alter the whole design. Substituting a processor requires engineering work. Changing wireless modules can trigger new certification. Replacing software infrastructure can affect performance.

Enterprise buyers face a related problem. A robot is not a one-time appliance when it depends on updates, remote management, spare parts, and vendor support.

Customers need assurance that deployed systems will remain serviceable across their expected lives. Policy instability makes those commitments harder to price.

This risk does not prove Unitree will lose its overseas business. The company has existing product authorizations, operations across multiple markets, and strong manufacturing capabilities.

It does mean historical international growth cannot be projected mechanically. Future models might encounter market-specific delays even when customer demand remains healthy.

The IPO proceeds could help Unitree build alternative technologies and expand domestic deployment. China’s manufacturing base offers suppliers, engineering talent, and government support for embodied AI.

However, domestic substitution does not immediately replace the strategic value of overseas revenue. International customers provide foreign currency, varied use cases, technical feedback, and brand legitimacy.

The policy risk also widens the gap between Unitree shares and synthetic contracts. An official stock market can process filings, trading suspensions, and material disclosures under established rules.

An offshore derivative venue might react through funding rates, margin changes, or contract adjustments. Traders could face forced liquidation before they can assess a new policy announcement.

This is why the underlying security matters. Price exposure without ownership can look efficient during stable conditions. During a regulatory shock, the contract’s legal and settlement details become decisive.

What the First Weeks of Trading Will Actually Reveal

Three signals will show whether Unitree’s market debut reflects durable confidence or a short-lived scarcity premium.

The first signal is the official listing announcement and opening-day market structure. Investors should rely on the Shanghai Stock Exchange, not a social post or derivative venue.

The exchange’s listing review record confirms Unitree’s regulatory path and planned use of proceeds. A separate official announcement must establish the trading date and final listing arrangements.

Once trading begins, the relevant information includes turnover, price range, and the balance between institutional and retail activity. A large opening move alone does not establish a stable valuation.

New A-share offerings often have limited freely tradable supply. Unitree’s offering also includes lockup arrangements for portions of institutional and strategic allocations.

Scarcity can amplify early moves in either direction. Investors should watch whether liquidity improves after the initial excitement and whether the stock finds support across several sessions.

The second signal is operating performance after the listing. Revenue growth must be evaluated alongside gross margin, research spending, and adjusted profit.

Unitree’s first-quarter results already showed the tension. Sales increased substantially, but adjusted earnings fell as costs and competition rose.

A stronger thesis requires evidence that deployments scale without proportionate increases in support and marketing expenses. Stable margins would suggest Unitree retains pricing or manufacturing advantages.

A weaker thesis would appear through repeated discounting, slower revenue growth, or declining profit despite larger shipments. Those results would indicate that market share is becoming more expensive to defend.

Customer composition will matter as well. Additional research and education demand supports volume, but diversified industrial adoption would strengthen the long-term case.

Investors should look for repeated factory, inspection, logistics, and service deployments. Named customers and measured operating results matter more than pilot announcements.

The third signal is the treatment of future Unitree products in overseas markets. Existing authorizations provide some continuity, but new models face greater uncertainty.

An exemption or conditional US approval would reduce immediate concerns. More restrictions, revoked permissions, or delayed launches would weaken the international growth story.

These policy outcomes also test Unitree’s product architecture. A company that can substitute components and certify regional models quickly has more resilience than one dependent on a fixed configuration.

None of these signals can be inferred reliably from a pre-IPO perpetual contract. Its price combines expectations about Unitree with leverage, venue liquidity, and platform-specific rules.

The contract might rise before the official listing and still settle poorly for individual traders. It might fall even while Unitree’s offering succeeds. Those outcomes are not contradictory.

The Unitree IPO deserves attention because it brings unusually detailed financial disclosure to humanoid robotics. It gives public investors a reference point for a sector dominated by private-company claims.

Yet the offering does not validate every adjacent speculation. Official shares, subscription rights, private secondary interests, and crypto derivatives remain distinct instruments.

Readers following the debut should keep a dated record of filings, product claims, policy changes, and deployment evidence. A personal second brain can preserve that chronology when headlines move faster than verification.

The practical question is simple: are you evaluating Unitree’s operating business, or trading a platform’s interpretation of future demand?

If the goal is business analysis, follow exchange filings, financial results, customer deployments, and regulatory decisions. If the goal is speculation, read every settlement and liquidation term first.

Unitree’s formal listing will eventually produce a real public-market price. Until then, a contract bearing its name remains a proxy, not privileged access to the stock.