The numbers are breathtaking. A subscription rate of 0.02% to 0.03% — orders of magnitude lower than the 0.47% seen for Changxin Technology’s recent IPO. The projected first-day return per lot: over 200,000 RMB based on historical averages of 276% for all STAR Market stocks and 466% for tech-focused ones. Yet, at the core of this frenzy lies a void. The company behind it — Unitree Technology, the so-called “first humanoid robot stock” on China’s Science and Technology Innovation Board — has not disclosed its price-to-earnings ratio, its revenue breakdown, or even the exact use of proceeds. The market is trading a narrative, not a balance sheet.
When the graph spikes, the soul remains quiet. This is not a crypto token launch; it is a traditional IPO built on the same religious fervor that once drove ICOs and DeFi liquidity mining. As a decentralized protocol PM who has witnessed the arc of blockchain from Gitcoin’s quadratic funding to the Terra collapse, I see Unitree’s listing as a perfect case study of how speculative infrastructure — whether centralized or decentralized — can decouple from fundamentals. Let me walk through the dimensions that matter, drawing on my own scars from building in crypto’s boom-and-bust cycles.
Context: The Rare Coin and the Discounted Cash Flow
Unitree Technology is a leading manufacturer of quadruped and bipedal robots. Its public products — the H1 and G1 humanoid robots — use self-developed frameless torque motors and planetary reducers, achieving a low-cost mass-production approach that starkly contrasts with Boston Dynamics’ hydraulic legacy. The company has shipped thousands of units for industrial inspection, firefighting, and research. But the humanoid robot space is still in pre-commercial demonstration: most orders are small-batch, and the “embodied intelligence” layer — the AI that enables generalized manipulation in unstructured environments — remains a known bottleneck across the industry.
Why does this matter? Because the IPO’s extreme scarcity of tradable shares (tiny float) and the resulting hype mirror the “fair launch” narratives of crypto projects. In DeFi, we saw liquidity mining programs that drove TVL to absurd heights, only to collapse when rewards dried up. Here, the low subscription rate is a deliberate design — a small float creates a price spike on day one, benefiting early institutional investors and venture capitalists who hold locked-up shares. The market is rewarding “first-mover” status, not technical maturity. This is the same mental model that pumped Dogecoin or Shiba Inu: scarcity of supply, velocity of narrative, and a herd of retail investors chasing a symbol of the future.
I recall a similar dynamic during the 2020 DeFi Summer, when Uniswap’s UNI token launch saw a 300% surge on day one. The protocol had real usage — over $1 billion in daily trading volume — but the token’s price was driven by finite supply and emotional demand, not discounted cash flows. Unitree’s IPO is the same playbook, but with a real company that has physical assets and employees. That makes it both more dangerous and more interesting.
Core: Deconstructing the IPO Through a Crypto Lens
Let me break down the analysis along five dimensions — technology, commercialization, competition, investment, and infrastructure — while drawing parallels to blockchain ecosystems.
Technology: The Hardware Layer vs. The AI Stack
Unitree’s strength lies in hardware engineering. Their quadruped robots have a global market share of over 60% at one point, and their humanoid robot G1 is priced at around 100,000 RMB — a fraction of Tesla’s Optimus or Boston Dynamics’ Atlas. This is a classic “low-cost volume” strategy, similar to how Ethereum’s layer-2 solutions like Arbitrum and Optimism compete on transaction fees. However, the real bottleneck is the “AI brain.” Unitree has not demonstrated a proprietary large language model or a robust reinforcement learning framework for generalized manipulation. In crypto terms, they have a fast block production layer but lack a smart contract execution environment that can handle complex logic.
From my Gitcoin days, I learned that open-source hardware can be commoditized quickly. The true moat in robotics will be the data flywheel — the more robots you deploy, the more real-world data you collect, the better your AI models become. But Unitree needs to scale its fleet to millions of units to achieve that. With a small float and a valuation that likely exceeds 50 billion RMB (implied by the IPO frenzy), the market is pricing in that data flywheel before it exists. This is the same mistake we saw with Terra’s algorithmic stablecoin: the belief that scale alone would solve the stability problem. It didn’t.
Commercialization: The Dual Structure of Revenue
The company operates a dual-track business: quadruped robots provide cash flow (industrial and consumer sales), while the humanoid robot segment is a narrative-driven R&D project. This is analogous to a DeFi protocol that has a stablecoin lending market (real revenue) and a governance token that is purely speculative. In crypto, we’ve seen how protocols like Aave have sustainable revenue, while others like OlympusDAO relied on a “3,3” narrative that eventually collapsed. Unitree’s IPO prospectus — if it ever becomes public — will be the key to determining whether the humanoid robot segment is a value creator or a value destroyer.
Based on my experience consulting for a DeFi protocol during the 2021 liquidity mining boom, I saw how teams would artificially inflate TVL with incentive programs. The “active users” were just mercenary capital. Unitree’s “orders” for humanoid robots may be similar: small-scale, proof-of-concept orders that do not represent recurring demand. The IPO’s expected first-day return is based on historical averages, not on a discounted cash flow model. That is a red flag for any long-term investor.
Competition: The VCs, The Tech Giants, and The Open Source Threat
Unitree is a small player in a global race. Tesla, with its Dojo supercomputer and Full Self-Driving AI, is years ahead in embodied intelligence. Boston Dynamics is backed by Hyundai and Toyota Research Institute. NVIDIA is building an entire ecosystem of simulation and AI tools for robotics. Unitree’s competitive advantage is cost and speed to market, similar to how Chinese phone manufacturers captured market share from Apple. But in the crypto space, we’ve seen how a single open-source breakthrough can erode a proprietary advantage. For example, the rise of zkSync and Starkware threatened Ethereum’s dominance by offering cheaper and faster transactions. Unitree’s hardware is not protected by strong patents — the company’s IP portfolio is unknown. If a tech giant like Tesla or a Chinese state-backed entity releases a cheaper robot, Unitree’s moat vanishes.
I remember during the 2022 bear market, I advised a startup that was building a decentralized identity protocol. We thought our unique encryption scheme was a moat until a team from ETHGlobal released a more efficient implementation as open-source code. The barrier to entry was low. Similarly, humanoid robot hardware is a commodity that can be reverse-engineered. The real value is in the AI and data, which Unitree has not yet demonstrated.
Investment: The SPAC-ification of the IPO
This IPO has all the hallmarks of a Special Purpose Acquisition Company (SPAC) deal: high uncertainty, low float, and a narrative-driven valuation. The subscription rate of 0.02% means that only a tiny fraction of retail investors will get shares, and those who do will likely sell on day one to capture the 276% average gain. The real volume will come from institutional investors who are allocated shares through the strategic placement. This is similar to how crypto exchanges like FTX used “initial exchange offerings” (IEOs) to create artificial scarcity. The price discovery is flawed because the float is too small to reflect true demand.
From my work on the Bitcoin ETF regulatory framework in 2025, I learned that liquidity is a function of transparency and trust. An IPO with a tiny float is like a token with a locked supply: it creates a false sense of value. The risk is that after the lockup period expires (usually 6-12 months), the share price can collapse. We saw this with several SPACs in 2021-2022, where the stock dropped 80% after the merger. Unitree’s IPO is a high-stakes gamble.
Infrastructure: The Compute Pipeline
Humanoid robots require massive compute for training and inference. Unitree likely uses off-the-shelf NVIDIA Jetson modules for edge AI and cloud GPU clusters for training. The cost of compute is a significant operational expense. In crypto, the equivalent is the gas fee for executing smart contracts. If Unitree’s AI training costs are too high, the company will bleed cash, just as a DeFi protocol with high gas fees loses users to cheaper alternatives. The IPO proceeds will likely be used to build a proprietary AI training cluster or to secure a partnership with a cloud provider. But the company has not disclosed this. Based on my experience managing a DeFi protocol’s treasury, I know that compute costs can easily eat into gross margins. Unitree’s ability to control that cost will determine its long-term profitability.
Contrarian: The True Value Might Be in a DAO, Not a Corporation
Here is the counter-intuitive angle: perhaps the most valuable application of humanoid robots is not a single company, but a decentralized autonomous organization (DAO) that owns a fleet of robots and rents them out to perform tasks. Imagine a robot network that is open-source, with owners pooling compute power and data to improve the AI collectively. This is the “robot sharing economy” — a decentralized physical infrastructure network (DePIN) like Helium or Filecoin, but for labor. In that model, Unitree’s IPO is a trap: it locks the value into a traditional corporate structure, subject to shareholder lawsuits, regulatory friction, and middlemen. The blockchain community has already shown that tokenized ownership can unlock liquidity and global participation.
I recall a conversation with a fellow builder during the 2023 bear market: we discussed how a robotics company could issue a token that represents a claim on the robot’s future earnings. The token would be traded on a decentralized exchange, and the robot’s task execution would be verified by a smart contract. This is not science fiction; it is already happening in the logistics sector with autonomous trucks. Unitree could have choosen that path, but they went the traditional IPO route because it is easier to navigate within China’s regulatory framework. However, by doing so, they are limiting their potential upside. The contrarian view is that the IPO is a signal of weakness, not strength.
Takeaway: The Robot’s Soul Is Still Quiet
Unitree’s IPO is a mirror of the crypto market’s most toxic behaviors: narrative over fundamentals, scarcity over utility, and short-term gain over long-term viability. The lock of the first-day return is a siren call that will lure retail investors into a speculative frenzy, but the underlying reality is a company with a nascent product, no visible AI moat, and a tiny float that will be violently volatile.
Based on my experience building in decentralized finance, I have learned that hype cycles are the enemy of sustainable value. The projects that survive are those that ship working products, engage with their communities authentically, and build infrastructure that can withstand market downturns. Unitree may succeed — their hardware is genuinely impressive, and the demand for humanoid robots is real. But the IPO structure means that the founders and early investors are cashing out first, leaving retail to hold the bag when the narrative shifts.
When the graph spikes, the soul remains quiet. The real test for Unitree will come not on the first day of trading, but when the lockup expires, the quarterly earnings disappoint, and the next shiny thing — a quantum computing IPO, perhaps — steals the spotlight. That is when we will see whether this robot company has a soul, or just a well-engineered chassis.