12.66 meters per second. That's the number Unitree wants you to remember. Faster than Usain Bolt. But the sprint isn't on the track; it's in the order book. The reveal landed Monday, days before the Shanghai IPO opens. Investors had already pushed the share sale to a record subscription level last week. Cold hands dissect the heat of a hype cycle.
Context: The IPO Record and the Robot Reveal
Unitree priced its shares at 150.8 yuan, raising roughly 6.1 billion yuan ($905 million) at a near $9 billion valuation. Retail buyers covered their tranche 8,288 times over. Shanghai's STAR Market had never seen such demand. The company sought 4.2 billion yuan and walked away with 6.1 billion—a 45% overshoot. The deal values Unitree at about 36 times its 2025 sales, against roughly 18 times for Hong Kong-listed rival UBTech. Chinese memory maker CXMT showed how violent these debuts get: its chip listing climbed 466% in one session on the same board.
Real revenue sits behind the story. Sales reached 1.7 billion yuan last year, more than four times the 2024 figure, while net profit hit 591 million yuan. Even so, buyers are paying above 100 times those earnings. The robot announcement—a humanoid called Superman claiming a top speed of 12.66 m/s—is timed to boost the narrative before trading begins.
Core: Systematic Teardown of the Speed Claim and Valuation
Let's start with the physics. Unitree says Superman clears a 2-meter standing high jump on legs measuring 0.85 meters. That's a jump height-to-leg ratio of 2.35. For context, a human with 0.85-meter legs would need to jump over 2 meters to match that ratio—possible, but only for elite athletes. The company claims the machine was built in a little over three months. That's a rapid development cycle, but speed doesn't equal reliability. No independent verification has been released. The kinematic analysis of Bolt's 2009 world record run published in the Journal of Experimental Biology pegged his peak speed at 12.42 m/s. Unitree's number clears that by 0.24 m/s—about 2% faster. In a sprint, that's margin of error territory. Without a third-party timing system and a certified track, the claim is a marketing number, not a scientific one.
I've traced enough robotic control systems to know that a sprint record on a flat track is not the same as reliable operation in a warehouse. 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 rather than factory floors. That's the real story: the company is selling toys for scientists, not tools for manufacturers. The IPO proceeds go toward embodied AI, new robot bodies, and factory capacity. But the current revenue base—1.7 billion yuan—is still tiny compared to the valuation. At 36x sales, you're betting on a 10x revenue expansion in the next few years. That's not impossible, but it's priced in as if it's already happened.
Now look at the demand. Retail investors oversubscribed their tranche 8,288 times. That's not a signal of deep fundamental analysis; it's a herd stampede. The STAR Market is known for these irrational pops. CXMT's 466% first-day surge is the benchmark. Unitree's IPO will likely open at a similar premium, then bleed as insiders cash out. The valuation is propped by hype, not by the cold numbers of industrial contracts. The company has no major factory deployment deals announced. Competitors like Tesla are building humanoid robots for their own factories. Unitree is selling to research labs. That's a different business model with different margins.
Contrarian: What the Bulls Got Right
But the bulls aren't entirely wrong. Unitree's revenue growth is real: 4x year-over-year. The net profit of 591 million yuan shows they can monetize the hype. The Superman robot, even if the speed claim is exaggerated, demonstrates engineering capability. The company built a humanoid that can jump and run in three months. That's a non-trivial achievement. Capital continues to chase robotics: Tether led a $1.4 billion round for NEURA Robotics in June. NVIDIA struck robotics deals with LG and Doosan the same month. Elon Musk is pouring billions into a record-sized chip factory. The narrative is that humanoid robots will replace human labor in factories and warehouses. Unitree is a pure play on that narrative, and the IPO demand reflects that belief.
However, the contrarian view is that the narrative is ahead of the technology. The robot's speed record is a one-off stunt. The real metric is durability, cost, and dexterity in unstructured environments. Unitree's robots are still experimental. The 5,500 units shipped are a drop in the ocean compared to the millions of industrial robots in use. The valuation assumes that Unitree will capture a significant share of that market. But the company has no proven track record of large-scale deployments. The IPO is a bet on the future, not a reflection of the present.
Takeaway: Accountability After the Bell
The sprint record is a distraction. The real question is whether Unitree can convert IPO capital into factory-floor contracts. We audit the code, but we mourn the users. In this case, the users are retail investors who will buy the IPO at the peak and watch it bleed. The company's claim of a 12.66 m/s robot is unverified. Independent auditors should be called in to test the robot's performance under controlled conditions. Until then, the speed is a marketing number, not a technical one. The IPO will likely pop on day one, but the long-term value depends on whether Unitree becomes a supplier to factories, not just a supplier to labs. The fork wasn't—the sprint wasn't the signal. The real signal is the order book. Watch that, not the video.
