Hook
Unitree claims a humanoid robot named Superman just ran 12.66 meters per second. Faster than Usain Bolt's peak. The reveal lands three days before the company's Shanghai IPO starts trading.
Investors have already pushed the share sale to a record subscription level. Retail buyers covered their tranche 8,288 times over. That is not demand. That is a stampede.
Silence in the ledger speaks louder than hype. The robot's speed record is unverified. The IPO valuation is 36 times sales. The revenue is real, but the multiples are fantasy. Let me walk through the code.
Context
Unitree priced its shares at 150.8 yuan, raising 6.1 billion yuan — roughly $905 million. The sale values the company near $9 billion. The company sought 4.2 billion yuan and walked away with 6.1 billion, a 45% overshoot. Shanghai's STAR Market had never seen demand like this.
Sales reached 1.7 billion yuan last year, more than four times 2024's figure. Net profit hit 591 million yuan. Even so, buyers are paying above 100 times those earnings. Hong Kong-listed rival UBTech trades at roughly 18 times sales. The gap is not a discount. It is a warning.
Superman is the headline. The robot clears a 2-meter standing high jump on 0.85-meter legs. Engineers built it in three months. Founder Wang Xingxing predicted in March that humanoid machines would break human sprint limits by mid-year, citing cheaper components and faster algorithms. Five months later, the claim arrived.
Core
Let me break down the technical claims. Bolt peaked at 12.42 meters per second during his 2009 world record run, according to published kinematic analysis. Unitree's number clears that by 0.24 m/s. That is a 1.9% margin.
But the company released no independent verification. No third-party lab. No video with calibrated timing gates. No data from onboard sensors timestamped against a certified reference. In my 2017 ICO audit days, I learned that silence in the technical appendix is a red flag. If the claim is real, show the raw telemetry.
Speed without structure is just noise. A robot that runs fast in a straight line is a different machine from one that navigates a factory floor with obstacles, tools, and human workers. Unitree shipped more than 5,500 humanoid units in 2025 across its G1, H1, and R1 lines. Most went to research labs and entertainment buyers. Not factory floors.
The revenue story is real, but the composition matters. 1.7 billion yuan in sales — impressive growth. But how much of that is from humanoid robots versus quadrupedal units or other products? Unitree's earlier robots like the Go1 and B2 are four-legged, and they generate steady revenue. The humanoid lines are still early-stage.
IPO proceeds go toward embodied artificial intelligence, new robot bodies, and factory capacity. That is a capital-intensive bet. The question is whether the market is pricing in a success scenario that assumes industrial adoption within two years. Based on my experience analyzing DeFi yield farms in 2020, I saw the same pattern: high multiples justified by hypothetical future cash flows, with no clear path to realization.
Data does not negotiate; it only confirms. The IPO demand is data. The subscription ratio is data. But the underlying business metrics — unit economics, customer retention, repeat orders — are still opaque. The 8,288x retail oversubscription is a signal of sentiment, not of fundamental value.
Contrarian
The contrarian angle is not that Unitree is a fraud. It is that the market is misreading the signal.
A sprint record does not equal industrial orders. The robot's speed is impressive, but industrial robotics requires precision, reliability, and safety certification. A 12.66 m/s robot that runs in a straight line is a marketing asset, not a factory asset. The real bottleneck for humanoid robots is not speed. It is manipulation, dexterity, and cost.
Yield is not income; it is risk repackaged. The IPO valuation of 36x sales is a yield on future growth. But the market is ignoring the structural risk: humanoid robots are still a niche product. The humanoid robot market is projected to be worth $38 billion by 2035, according to some estimates. That is a decade away. Unitree's current valuation implies they will capture a significant share of that market very quickly.
Compare to NEURA Robotics, which raised $1.4 billion from Tether in June. NVIDIA struck robotics deals with LG and Doosan the same month. Elon Musk is pouring billions into a record-sized chip factory. Capital is flowing into the sector, but that does not mean every company will win.
The audit trail never lies, only the auditor can. The lack of independent verification for the Superman speed record is a small piece of a larger pattern. Unitree has not released a detailed technical paper on the robot's control system, nor have they disclosed the battery life, thermal management, or failure rates. These are the metrics that matter for industrial deployment.
Takeaway
Watch the IPO trading debut. If the stock opens at a 50% premium or higher, the market is pricing in a fantasy. If it opens modestly, the market may be more rational than the subscription ratio suggests.
But the real signal will come in six months. Unitree needs to convert the hype into industrial orders. If they announce a partnership with a major manufacturer, the sprint record will have been a catalyst. If they announce another robot with a higher jump, the market will realize the distraction.
Speed without structure is just noise. The question is whether Unitree has the structure to match the speed. Based on the data so far, I am skeptical. The ledger does not lie. The silence is telling.