Unitree revealed a humanoid robot called Superman on Monday, claiming a top speed of 12.66 meters per second. That pace edges past the fastest stride Usain Bolt ever recorded. The reveal lands days before the Unitree IPO starts trading in Shanghai. Investors had already pushed the share sale to a record subscription level last week.
Hook
A machine that outruns the fastest human ever measured. A retail subscription ratio of 8,288 times over. Unitree’s IPO on Shanghai’s STAR Market just set a record for demand. The company raised $905 million at a valuation near $9 billion. That is 36 times its 2025 sales. The number screams speculative frenzy. But behind the noise sits real revenue: 1.7 billion yuan last year, net profit 591 million. The market is pricing in a future where humanoid robots replace human labor at scale. Verify the proof, ignore the hype.
Context
Unitree is a Chinese robotics company founded in 2016 by Wang Xingxing. It started with quadruped robots (the Laikago series) and moved into humanoids in 2023 with the H1. The current lineup includes the G1, H1, and R1. The company shipped over 5,500 humanoid units in 2025, mostly to research labs and entertainment buyers. The IPO on the STAR Market (Shanghai’s Nasdaq-style tech board) is the largest robotics IPO in China this year. The proceeds will fund embodied AI, new robot bodies, and factory capacity.
Superman, the new robot, stands on legs 0.85 meters long. Unitree claims it can clear a 2-meter standing high jump and sprint at 12.66 m/s. That beats Usain Bolt’s peak speed of 12.42 m/s during his 2009 world record 100m. The company released no independent verification. Founder Wang had 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 lands.
Core Analysis: The Numbers Behind the Robot
Let’s dissect the financials. Unitree’s revenue grew 4x in 2025 to 1.7 billion yuan. Net profit hit 591 million yuan, implying a net margin of 34.8%. That is impressive for a hardware company. But the IPO values the company at 36x sales. For context, UBTech, a Hong Kong-listed rival, trades at about 18x sales. The premium is 2x. The price-to-earnings ratio? Over 100x. Code is law, but bugs are reality. The valuation is betting on exponential growth.
Retail buyers covered their tranche 8,288 times over. That is the highest ever on the STAR Market, beating the previous record set by CXMT (a memory chip maker) which saw a 466% first-day pop. Unitree’s IPO was upsized from 4.2 billion yuan to 6.1 billion yuan, a 45% overshoot. The demand is reminiscent of the 2021 NFT mania in crypto—where token sales were oversubscribed by thousands of times based on a narrative, not fundamentals.
But is there substance? 5,500 humanoid units shipped. Assuming an average selling price of $50,000 per unit (rough estimate for H1 and G1), that gives $275 million in robot revenue. The rest of the 1.7 billion yuan (about $240 million) likely comes from quadruped robots, components, and services. The market is valuing each robot sold at about $1.6 million in market cap. That is a high multiple per unit.
Compare to Tesla’s Optimus. Tesla has not shipped humanoids at scale yet. NVIDIA is selling chips and software for robotics, not the robots themselves. Unitree is actually producing and selling. That is a real asset. But the question is: can they scale production to tens of thousands of units per year? The IPO proceeds will go toward factory capacity. A robot factory is capital-intensive. The $905 million raised will help, but it is not enough to build a Gigafactory.
Superman’s performance numbers are impressive on paper. 12.66 m/s is 45.6 km/h. The world record for a 100m sprint is 37.6 km/h. So the robot is 21% faster than the fastest human. But there is no verification. Unitree released a video on Twitter, but no third-party timing or certification. In the crypto world, we call that a white paper without a working demo. Optimism is a feature, not a guarantee.
The robot’s legs are 0.85 meters long. For a human, that would be a tall person. The standing high jump of 2 meters is also impressive—roughly the height of a basketball hoop. But jumping and running are different skills. The robot likely uses a combination of high-torque motors and advanced control algorithms. The development time of just over three months suggests Unitree is iterating fast. But fast iteration can lead to bugs. In my experience auditing smart contracts, speed often comes at the cost of security. The same applies to hardware.
Let’s talk about the embodied AI angle. Unitree says the IPO proceeds will go toward embodied AI. That means software that controls physical machines. This is the intersection of AI and robotics. The market is hot. NVIDIA signed robotics deals with LG and Doosan. Tether led a $1.4 billion round for NEURA Robotics. Elon Musk is building a massive chip factory. Unitree is positioning itself as a pure-play humanoid robot company. But the competition is fierce. Tesla, Boston Dynamics, Figure, Agility Robotics, and numerous Chinese startups are all chasing the same prize.
Unitree’s advantage is that it already has revenue. Most humanoid robot companies have zero revenue. Figure is still in prototype stage. Boston Dynamics has been sold twice. Unitree is shipping. That is real. But the revenue is from research labs and entertainment, not industrial applications. Factory floors are the holy grail. Can Unitree convert its sprint record into industrial orders? That is the key question.
Contrarian Angle: The IPO Valuation Is a Security Blind Spot
Let me be contrarian. The IPO demand is a function of the Chinese retail investor frenzy, not a sober assessment of the company’s prospects. The STAR Market is known for its volatility. CXMT rose 466% on its first day, then corrected. Unitree’s valuation at 36x sales is already pricing in years of growth. If the company grows revenue at 100% per year for the next three years, it would still be trading at 18x future sales. That is not cheap. UBTech, a comparable, trades at 18x current sales. The premium for Unitree is 2x. Is Unitree twice as good as UBTech? UBTech also has humanoid robots, but its sales are lower. Still, the market is giving Unitree a huge benefit of the doubt.
Superman’s speed record is unverified. In the world of blockchain, we call that a “trust me bro” claim. Unitree has not released a peer-reviewed kinematic analysis like the one used for Bolt. The company has a history of impressive demos, but we need to see the robot run in a controlled environment with independent timekeepers. The same applies to the high jump. Without verification, the claim is just marketing.
Another blind spot: the cost of the robot. If Superman costs $200,000, it is not competitive with human labor. A factory worker in China costs $10,000 per year. The robot would need to work for 20 years to break even. Battery life, maintenance, and reliability are unknown. Unitree’s existing robots are used in labs, not factories. Industrial robots require 24/7 operation with minimal downtime. Humanoid robots are still far from that.
Finally, the IPO itself is a liquidity event for early investors. The founder and early backers will likely sell some shares. The lock-up period is typically 6 months for the STAR Market. After that, the market will absorb the supply. The retail buyers who oversubscribed at 8,288x are likely to flip the shares on the first day for a quick profit. That creates volatility. The long-term holders are the ones who believe in the vision. But the valuation is already stretched.
Takeaway: The Robot Is Real, the Valuation Is Speculative
Unitree is a real company with real revenue and real robots. Superman is a real technical achievement. But the IPO valuation is a bet on the future of humanoid robotics, not a reflection of current earnings. The stock will likely be volatile. The first-day pop could be huge, followed by a correction. The key metric to watch is industrial adoption. If Unitree can secure a contract with a major manufacturer like Foxconn or BYD, the thesis holds. If not, the robot will remain a lab curiosity.
In the crypto world, we have seen similar hype cycles. DeFi protocols with $1 million in revenue were valued at $1 billion. Most crashed. The ones that survived had real users and fees. Unitree has a chance to be the survivor. But the price is high. I would wait for the first quarterly report after the IPO to see revenue growth and robot shipments. Until then, Trust the math, not the roadmap.