Unitree’s IPO: A 0.02% Subscription Rate Hides the Real Code

The subscription rate was 0.02%. Not a typo. For Unitree Technology, the humanoid robot company listing on Shanghai’s STAR Market, only three out of every ten thousand applications got an allocation. The brokerage memo I read called it a ‘locked-in profit’ scenario: every signed share could net 200,000 RMB in first-day gains. Code doesn’t lie. Neither does the math of supply and demand. But this isn’t a token launch. It’s an IPO—and the same mechanics that made DeFi summer hot are now repackaged for A-share retail investors. Context matters. Unitree is not a blockchain project. It builds four-legged and bipedal robots. Their H1 and G1 models are priced around $15,000, a fraction of Tesla’s Optimus or Boston Dynamics’ Atlas. The company has real hardware: self-developed torque motors, planetary gearboxes, a closed-loop supply chain. They’ve shipped thousands of units to labs, fire departments, and hobbyists. But the IPO is not about hardware. The offering size is minuscule relative to market demand. The brokerage report I dissected projects a 0.02% to 0.03% allocation rate—far below ChangXin Memory’s 0.47% earlier this year. That’s the hook: a tiny float plus a ‘first humanoid robot stock’ narrative creates a scarcity premium. Retail investors are bidding for a piece of a story, not a balance sheet. Now, the core analysis. I’ve audited over 50 ICO smart contracts. The pattern here is identical. A small circulating supply, outsized media attention, and a projected first-day return of 276% (A-share average) or 466% (STAR Market average). The brokerage uses historical averages to estimate single-sign profit of 200,000+ RMB. But historical averages are not guarantees. In crypto, we call this the ‘high APY but no real users’ trap. The IPO’s small float is a deliberate design choice. It ensures high subscription demand, but it also amplifies volatility. On day one, a small number of shares changing hands can swing the price 50%. The same mechanics that pump a low-cap token on Binance. The underlying business? Unitree’s revenue from quadrupeds is real but modest. Their humanoid line is still in pilot delivery. The real valuation—if you could see it—would be a multiple of annual sales, not profits. The P/E ratio is not disclosed in the prospectus, which is a red flag. In my experience, when a company hides the valuation metric, the price is set by sentiment, not fundamentals. Here is the contrarian angle. The euphoria hides a fundamental blind spot: Unitree’s AI stack is weak. Their robots can run and jump, but they lack the general-purpose reasoning needed for unstructured tasks. The demos are choreographed. The company does not have a self-trained large model. They rely on external chips and frameworks. This is the equivalent of a Layer-2 sequencer with no fraud proof deployed. The code is not there. The ‘AI brain’ is the bottleneck. Tesla’s Optimus team has FSD data and Dojo supercomputers. Unitree has none of that. The IPO proceeds will go to production lines and algorithm R&D, but catching up to the tech giants is a multi-year journey. Meanwhile, the small float attracts short-term speculators, not long-term holders. The first-day pump will be followed by a grind as sellers exit. I’ve seen this movie in 2017 with ICOs that had a great website but no working product. The same pattern repeats. Takeaway: watch the first quarterly report. If Unitree’s humanoid shipments stay below 1,000 units, the narrative will crack. The 0.02% subscription rate is a signal of desperation, not value. The smart money is already in the lock-up tranches. Retail investors are buying a lottery ticket. Code doesn’t lie. The code of this IPO is a small float, hidden valuation, and a technology gap that no amount of hype can fix. The real question is not whether the first day is green. It’s whether the company can deliver a general-purpose robot in 2027. Until then, it’s just a story with a low supply.