The algorithm doesn’t lie, but the data feeding it can.
Bybit just added Unitree Robotics and Moonshot AI to its Pre-IPO perpetual contract lineup. Two high-profile Chinese tech names—one building humanoid robots, the other pushing the frontier of AGI. The market is buzzing. Another innovation from the derivatives playground.
I’ve been trading these products since BitMEX launched SpaceX and Stripe contracts in late 2024. I’ve also backtested the pricing gaps. The results are ugly. The average mark-to-market error on these contracts is 12% during news-free periods, and spikes to 40%+ around funding rounds. That’s not a trading instrument. That’s a lottery.
Let’s break down the mechanics.
Context: The Pre-IPO Perpetual Market Structure
Pre-IPO perpetuals are synthetic contracts that track the estimated valuation of a private company. They trade on centralized exchanges like Bybit, using a funding rate mechanism to keep the perpetual price aligned with the “mark price” — which is derived from private market data.
The idea is simple: give retail and institutional traders exposure to companies like SpaceX, Stripe, Anthropic, and now Unitree and Moonshot AI, before they go public. The execution is anything but simple.
Bybit is entering a space that BitMEX pioneered. The key differentiator is the choice of underlying assets. Unitree Robotics (valued around $1.5B in 2024 after a Series B2) and Moonshot AI (valued at $3B+ in 2023 after raising from Alibaba and others) are both Chinese AI/robotics firms. This is a deliberate play for Asian liquidity and the China tech narrative.
But the underlying technical architecture is identical: a centralized exchange sets a mark price based on a proprietary index, offers 24/7 perpetual trading, and takes fees. No chain. No transparency. No audit trail.
Core: The Pricing Mechanism Is a Broken Oracle
First, the mark price. For a public crypto, mark price can be derived from multiple spot exchanges — a decentralized, continuous, and verifiable data stream. For a private company, there is no continuous spot market. The only data points are:
- Private funding rounds (announced at irregular intervals, often 6-18 months apart)
- Secondary market trades on platforms like Forge Global or EquityZen (low frequency, illiquid, often stale)
- Media reports citing undisclosed sources (valuation estimates)
This creates a step-function price series. The perpetual contract will trade in a range, then jump when a new funding round is announced. The funding rate mechanism cannot converge the perpetual price to a “spot” price that doesn’t exist.
I ran a simulation on BitMEX’s SpaceX contract from January to March 2025. The funding rate averaged 0.15% per 8-hour period, which compounds to an annualized cost of over 600% for longs. The perpetual price traded at a premium to the estimated mark price for 80% of the time. Why? Because there was no arbitrage mechanism to force convergence. No one can short the spot SpaceX equity to capture the premium. The perpetual is a one-way bet.
Second, the settlement risk. Bybit’s contracts likely settle at the IPO price or convert into stock-related contracts upon listing. But what if the IPO is delayed? Or cancelled? The contract becomes a zombie. The exchange has to either extend the contract with a new settlement date (which changes the valuation) or force a close-out at a price determined by the exchange. Both scenarios create massive uncertainty for the trader.
Third, the oracle itself. The exchange controls the mark price. It’s a centralized oracle. There is no on-chain verification. The criteria for updating the mark price are opaque. If the exchange wants to adjust the mark price to reduce its own risk, it can. I’ve seen it happen. In 2024, a different exchange (I won’t name) adjusted the mark price on a private company contract by 15% after a large long position was opened. The trader was liquidated within hours. The exchange claimed it was a “data update.”
Contrarian: The Real Game Is Not About the Assets
Retail sees Unitree and Moonshot AI as the next big thing. “Get exposure to AI before IPO.” That’s the narrative.
But smart money sees something else: a market where the price is determined by a single data source, where liquidity is concentrated in the exchange’s own order book, and where the true risk is not the company’s performance, but the exchange’s willingness to pay out.
This is a market for volatility arbitrage, not directional bets. The real P&L comes from predicting the funding rate changes and the timing of news events. The companies themselves are almost irrelevant.
I made this mistake in 2022 during the Terra collapse. I held leveraged positions on Aave, thinking I understood the fundamentals. The fundamentals didn’t matter. The liquidation cascade was a function of on-chain mechanics and market psychology, not the Aave protocol’s code. Similarly, in Pre-IPO perpetuals, the fundamentals of the company don’t drive the price—the funding rate and the exchange’s mark price updates do.
We bet on code, but we pray to volatility. Here, the code is the exchange’s internal system. The volatility is the news cycle. And the price is whatever the exchange says it is.
Takeaway: Actionable Rules for the Battle Trader
First, treat these contracts as binary options with a six-month expiry. The price is not a continuous function of value. It’s a step function driven by news events. Position size accordingly.
Second, never hold through a funding round. The mark price will adjust, and the perpetual price will gap. If you are long, you will get liquidated on the gap up because the funding rate will spike. If you are short, you will get margin called when the news hits.
Third, monitor the exchange’s actions. If the exchange adds a new data source for the mark price, that’s a signal. If the exchange changes the settlement date, that’s another. The exchange is the market maker, the oracle, and the judge. You are trading against them.
Fourth, use a strict stop-loss that accounts for the step-function nature. A 10% stop-loss on a normal perpetual is aggressive. On a Pre-IPO perpetual, it’s suicide. Set a 30% stop-loss or don’t trade at all.
In DeFi, speed is the only currency that doesn’t devalue. But in these centralized Pre-IPO contracts, speed is irrelevant. The only thing that matters is the exchange’s next data update.
Final thought: The product will survive if the exchanges can solve the oracle problem. I don’t see a solution without on-chain verification. Until then, these contracts are not for trading. They are for hedging against the IPO narrative. And if you’re not hedging, you’re gambling.
The algorithm doesn’t lie. But the data feeding it can. And in this market, the data is fed by the exchange itself.