Hook
On August 19, 2025, a Chinese robotics company named Unitree went public on the STAR Market (the Nasdaq-style board of the Shanghai Stock Exchange) with an opening price of 1,100 yuan per share — a 629% spike from its IPO price of 150.8 yuan. The market cap hit 444.9 billion yuan (about $62 billion). That’s larger than the combined market cap of every single AI token on CoinMarketCap today. One early investor, Astrend IV (a vehicle of Lei Jun’s Shunwei Capital), saw a paper gain of 15.2 billion yuan. The news didn’t just shake the A-share market — it sent a shockwave through the crypto AI narrative, where tokens like Render, Bittensor, and Akash have been pricing in similar “physical world AI” fantasies.
Context
Unitree is the first publicly listed company from the so-called “Hangzhou Six Little Dragons” — a group of AI and robotics startups backed by local government policy. The company is best known for its quadruped robots (Go2, B2) and humanoid robots (H1, G1). Unlike many AI startups that only raise money, Unitree has actually shipped products. It has a commercial loop: consumer-grade quadrupeds sold globally, industrial-grade units deployed in power inspection and security, and now humanoid robots priced aggressively at 99,000 yuan. This is rare in the global robotics industry. But the market’s 444.9 billion yuan valuation is not about the current business. It’s an option on the future of humanoid robots. In crypto terms, it’s like buying a token with no revenue but a strong narrative and a pending roadmap. The difference is that Unitree has real hardware, real supply chains, and real early revenue. Yet the price-to-sales multiple implied by the IPO is extreme — even by crypto standards. In the crypto AI sector, projects like Bittensor (TAO) trade at a market cap of ~$3 billion with $0 in direct revenue from the network. Unitree’s $62 billion valuation against an estimated <$2 billion revenue puts it in a similar territory of narrative-driven pricing. The key question: does Unitree’s IPO validate or invalidate the crypto AI narrative?
Core
I’ve been tracking the convergence of AI and crypto since 2020, and I can tell you that Unitree is the first “real-world” proof that the market is willing to pay a massive premium for companies that execute on the physical AI dream. This has direct implications for crypto AI tokens. Let me break it down with three data-driven insights.
1. The valuation anchor effect
Unitree’s IPO sets a new floor for how much capital markets are willing to allocate to “embodied intelligence.” Before Unitree, the largest public AI company in China was SenseTime (market cap ~$10 billion). Unitree, with a $62 billion market cap, is now 6x larger. This means that the public market is now pricing hardware AI execution at a premium over pure software AI. In crypto, this is a direct signal for DePIN (Decentralized Physical Infrastructure Networks) projects like Helium, Hivemapper, and Filecoin, which also combine hardware with token incentives. If Unitree can command a $62 billion valuation for a centralized robot company, what is the ceiling for a decentralized network of physical sensors or compute nodes? The crypto market has been struggling to find a valuation anchor for DePIN — Unitree just provided one. But note: Unitree’s valuation is based on actual product sales and a clear path to humanoid mass production. Most DePIN projects have negligible revenue. The gap is huge, but the narrative pull is strong.
2. The capital flow redistribution
Shunwei Capital’s 15.2 billion yuan paper profit represents a 60x+ return on its early investment. This will trigger a herd effect in the venture capital world. Expect to see more money flowing into “hardware + AI” companies, both in traditional markets and in crypto. Crypto AI tokens that have a clear hardware component — like Akash (compute), Render (GPU rendering), and Bittensor (mining compute) — will benefit from the narrative spillover. But I’ve seen this pattern before. In 2021, when Coinbase went public at a $100 billion valuation, it boosted the entire crypto exchange narrative. Yet the actual correlation between Coinbase’s stock and altcoin prices was short-lived. The real money went to the underlying infrastructure. Similarly, Unitree’s IPO will likely boost the AI token sector in the short term, but the long-term winners will be crypto projects that have real hardware deployment and revenue, not just speculation.
3. The narrative convergence
Unitree’s narrative is about “mass production humanoid robots as the next computing platform.” This is exactly the same story that AI tokens are selling: “decentralized AI compute as the foundation for AGI.” Both narratives rely on a future where AI agents need physical or virtual compute at scale. But Unitree has the hardware execution; crypto AI has the software tokenization of execution. The market is now pricing both with similar multiples. This is a classic “narrative merger” moment. I remember in 2020 when DeFi tokens like COMP and AAVE started to trade at multiples similar to traditional fintech IPOs. The market was saying, “DeFi is the new fintech.” Now, the market is saying, “Embodied AI is the new tech platform.” The crypto AI tokens that survive will be those that deliver actual compute or data, not just memes.
Contrarian
But here’s the contrarian angle that most retail investors are missing: Unitree’s IPO is not a bull case for crypto AI — it’s a warning. The 629% first-day pop is a classic sign of IPO underpricing by the investment bank to ensure a “hot” debut. It also reflects a massive gap between the institutional price (150.8 yuan) and the retail frenzy price (1,100 yuan). This gap is a bubble signal. In crypto, we see similar gaps when a token launches on a centralized exchange with a low initial market cap and then skyrockets on retail FOMO. The difference is that crypto tokens have no earnings, no disclosure, and no lock-up period. Unitree has a 1-3 year lock-up for early investors, which means the 152 billion yuan paper profit is not realizable for years. If the stock price falls during the lock-up, the paper gain evaporates. In crypto, early investors can dump immediately. So Unitree’s IPO is actually a more controlled bubble than crypto AI tokens. The real risk for crypto AI is that Unitree’s valuation sets an unrealistic expectation. If Unitree’s stock corrects by 50% in the next six months (which is common for STAR Market IPOs), the entire AI narrative in both traditional and crypto markets will take a hit. I’ve seen this happen with the “AI coin” mania in 2024, where tokens like Worldcoin dropped 80% after the initial hype. The lesson is: do not buy the narrative at the peak. Unitree’s IPO is the peak of the narrative, not the starting point.
Takeaway
The next narrative is not about humanoid robots themselves — it’s about the supply chain that enables them. In crypto, that means investing in decentralized compute networks (Akash, Render), sensor data marketplaces (Hivemapper, DIMO), and AI agent infrastructure (Bittensor). Unitree’s IPO validates the thesis that “hardware execution + AI = massive value.” But the market is already pricing that in. The real opportunity is in the “picks and shovels” — the decentralized infrastructure that will power the next generation of robots. As I always say, “Where code meets culture, the real value emerges.” Today, the code is Unitree’s hardware, and the culture is the market’s belief in AI. Tomorrow, the code will be the blockchain protocols that let robots communicate and transact. I’m watching.
