Unitree Doesn't Owe You 291%: The Pre-IPO Perp Is a Contract, Not a Promise

Most people see a number like 219,600 yuan per lot and stop reading. That is the first mistake. Unitree Robotics opens its STAR Market IPO subscription tomorrow, and Trade.xyz's pre-IPO perpetual contract is quoted at $87.525—roughly 590 yuan per share, or 3.91 times the official issue price. The arithmetic is seductive: subscribe 500 shares for 75,400 yuan, assume an exit at 590 yuan, collect 295,000 yuan, and walk away with 219,600 yuan of "guaranteed" profit. A 291% return before the company even trades. Logic doesn't care how loudly the robot dances. It asks a dirtier question: what are you actually buying? The answer, here, is a synthetic contract with no delivery obligation, settlement terms that are not spelled out in the parsed data, and a quote that may be real—or may just be the last thing a screen happened to display. Context first. Unitree has become one of the most recognizable names in quadruped and humanoid robotics. Its STAR Market IPO will issue 40.4464 million new shares, exactly 10% of the total post-issuance share count of about 404 million shares. At the official issuance price of 150.8 yuan, the offering looks conservative: roughly 6.1 billion yuan raised, with an implied valuation near 61 billion yuan. But the pre-IPO perpetual on Trade.xyz prices the same company at $87.525 per share. Converted at about 6.74 yuan per dollar, that's 590 yuan, and the implied post-issuance market cap jumps to approximately 238.7 billion yuan. That is not a small premium. It is a statement that the first-day close will be nearly four times the IPO price. The subscription mechanics are equally mechanical: one lot is 500 shares, a 75,400 yuan cash lock-up per lot, and if you are allocated and can sell at the perpetual's quoted price, the gross profit per lot is 219,600 yuan. Read the code, ignore the roadmap. The Unitree pitch deck is full of engineering beauty, but the pre-IPO perp is not equity. It is a cash-settled derivative, and its settlement math is the only math that matters. The 291% figure assumes the perpetual price is the future stock price. That assumption is far from automatic. First, pre-IPO perpetuals do not deliver shares. They settle against a reference price—often the first-day opening auction, sometimes a first-day VWAP, sometimes a price determined by an oracle with its own latency problems. The parsed materials do not state which reference applies to this contract. If settlement uses an opening auction that prints at 200 yuan instead of 590 yuan, the perpetual collapses to roughly $29.70. Your "291% return" becomes a 66% loss. The gap is not hypothetical. STAR Market IPOs have opened below the official issue price before. The issue price of 150.8 yuan is not a floor. Second, "last reported" is not a tradeable price. In a pre-IPO market without continuous regulated trading, the spread can be wide, the book can be thin, and the last print can be a wash trade or a stale quote from hours ago. During my 2025 audit of an AI-backed tokenized equity project, I traced a healthy-looking pre-IPO order book to three connected wallets that were feeding each other quotes. The screen showed momentum. The code showed circularity. Before treating $87.525 as the true market consensus, ask who supplied that price and how many shares could actually be purchased at that level. Third, look at the cost of carry. A perpetual contract carries a funding rate—a periodic payment from one side to the other based on how far the mark price drifts from the underlying. Here, the mark price is a 3.91x premium over the issuance price. That means the funding rate is likely to bleed long positions heavily. If the listing is delayed by two weeks, or the project faces administrative hurdles, the funding payments alone can erode the headline profit. The 291% is a gross, pre-funding, pre-fee, no-slippage fantasy. The fees on Trade.xyz are not zero. The funding rate is not zero. The variance is not zero. Fourth, the assumption that allotment is guaranteed is nonsense. The offering is only 10% of total shares. Oversubscription on STAR Market IPOs is the norm. If you subscribe for one lot and are not allocated shares, your cash is locked for days and returned with zero interest. The expected value of a 291% upside must be multiplied by the probability of receiving an allocation. That probability is often closer to a lottery ticket than an entitlement. During the DeFi summer audit era, I saw the same mistake repeatedly: people treated gross yield as if it were risk-adjusted profit, ignoring the capital lock-up, the failure probability, and the smart contract risk between the intended return and the delivered return. The core insight, then, is simple: the 291% return is not an arbitrage. It is a forecast of how much retail will overpay during the opening auction. The contract is a prediction market dressed up as an IPO subscription. The real question is not "will Unitree be worth more than 150.8 yuan?" The real question is "who is holding the other side of your synthetic long when the price prints below 590 yuan?" Now for the contrarian angle. The bulls are not entirely wrong. Unitree is a serious robotics company, and STAR Market quality issuers are scarce. Chinese IPO pricing tends to leave first-day money on the table to reward early investors, so the gap between the official issue price and the first-day trade price can be intentionally large. With a tight 10% float and strong retail demand, the opening auction may very well exceed the pre-IPO perpetual's implied price. In that world, buying the perpetual was a rational convex bet. Volatility is just unpriced risk, and a 3.91x premium can be a fair gamma price if you plan to exit before the perpetual converges to the spot market. The problem is not that the contract exists; it is that the settlement terms are not legally pinned down in the public material. You are not buying the future. You are buying a dispute about the future. Before you subscribe tomorrow, read the settlement specification on Trade.xyz. Ask whether the contract settles on the first trade, the first-day close, or a multi-day VWAP. Ask who decides the oracle price and what happens if the listing is delayed or canceled. Ask whether the funding rate is paid by the long side, and how much it has been in the first 72 hours. If you cannot answer those questions, you are not earning 291%. You are paying premium to someone who has already read the code. Logic doesn't leave money on the table unless the counterparty leaves the table first. The correct price for Unitree will become public after the opening auction. Do you want to be the one who pays for that discovery?