¥150.80. 500 shares per lot — ¥75,400 due per winning subscription.

On August 10, Unitree Robotics (688836.SH), "the first humanoid robot stock," officially opened STAR Market subscriptions. The issue price is ¥150.80 per share, with approximately 405 million total shares post-issuance, corresponding to a market cap of roughly ¥60.993 billion at listing. The trailing PE ratio is 219.23x — in the 2026 A-share market, that valuation level is enough to make most traditional industry analysts gasp.

But in the embodied AI space, 219x PE might just be the beginning of the beginning. After all, this is the first pure-play humanoid robot company on the A-share market — capital markets have been waiting for this ticker for a very long time.

Is ¥61B Expensive? Crunching on Three Dimensions

To discuss whether 219x PE is expensive, you first have to understand what kind of company Unitree actually is.

Many people assume Unitree is a humanoid robot company, but currently its primary revenue source is quadruped robots — those robot dogs you see everywhere. Humanoid robots are the future direction and the capital markets story, but they still account for a relatively limited share of current revenue. The company's product lineup includes industrial quadrupeds, consumer quadrupeds, humanoid robots, and robotic arms, with customers across research & education, industrial inspection, film & entertainment, and other sectors.

Evaluate the ¥61B valuation on three dimensions:

Dimension one: price-to-sales (PS) ratio. Based on 2025 revenue, the PS ratio is in the several-dozen range — much more expensive than traditional manufacturing, but not so outrageous compared to AI software companies. After all, robotics is a hardware business with supply chain costs; gross margins can never compete with pure software.

Dimension two: industry growth rate. Embodied AI is widely regarded as the next trillion-dollar track after AI; 2026 is called "year one of humanoid robot mass production" by many. If the industry truly achieves exponential growth over the next 3–5 years, 219x PE might look like a bargain in hindsight. But the question is: how probable is that "if"?

Dimension three: scarcity premium. There are almost no pure-play humanoid robot stocks on the A-share market. If capital wants exposure to the embodied AI track, Unitree is practically the only option. Scarce things are always expensive — the A-share market has always rewarded first-of-a-kind stocks with scarcity premiums. Previous examples like Cambricon and SenseTime enjoyed extremely high valuations early on. Whether it can be sustained depends on subsequent earnings delivery.

Why Embodied AI Commercialization Is Harder Than AI Software

Unlike "asset-light" tracks like foundation models and AI software, commercializing embodied AI is much harder.

It's hard because of hardware. A humanoid robot has dozens of motors, hundreds of sensors, thousands of components — supply chain complexity far exceeds that of an app. Every cost reduction and every yield improvement requires massive investments of time and money in supply chains, processes, and quality control. LLMs can iterate versions quickly; changing a single structural part on a robot might require months of tooling cycles.

It's hard because of use cases. Foundation models find value in a single use case — writing code, writing copy — but the killer app for humanoid robots hasn't been determined yet. Industrial inspection? Home service? Elderly care? Education and entertainment? Every direction has demand, but the magnitude of demand and willingness to pay in each are still being validated. Without a clear "must-have scenario," it's hard to achieve scale shipments.

It's hard because of competition. Too many companies are building humanoid robots now — overseas there's Tesla Optimus, Figure, Boston Dynamics; domestically there's UBTech, Fourier, Galaxy Intelligence, Zhiyuan Robotics — plus traditional automakers and home appliance giants all jumping in. Unitree may have listed early, but how high its technical moat is and whether its first-mover advantage can hold are open questions. And with the capital boost from going public, competitors will get more aggressive too — this track is going to get more and more crowded.

That said, the significance of being the first stock is never just about earnings. The real value of Unitree's IPO is providing a valuation anchor and exit channel for the entire embodied AI track. Primary market investors finally have a public-market benchmark; founders finally have a clear target; companies up and down the supply chain will be energized. It's just like new energy vehicles around 2020 — the first wave of listed companies might not be the ones that survive in the end, but they ignite the capital boom for the entire sector.

Is Unitree at 219x PE a bubble or hope? Nobody has a definitive answer right now. But one thing is certain: starting from the August 10 subscription date, embodied AI is no longer just lab demos and founder stories — it's become a real, tradable asset in the capital markets.

Time will tell whether it lives up to the hype.

See you tomorrow.