Unitree Is About to Become the First Pure-Play Humanoid Stock

The humanoid-robot industry has one shared assumption underneath every pitch deck: that these machines are a decade-out bet, a capital furnace you feed now against a payoff that arrives when the actuators get cheap and the models get good. It is the assumption that justifies the valuations and excuses the losses. Unitree Robotics, cleared in early July for a Shanghai STAR Market listing and finalizing pricing for a debut as early as late July, is about to test that assumption in public — because Unitree is profitable, and it is about to become the first pure-play humanoid company that anyone can buy.

The numbers that break the pattern

The offering is sized to raise roughly 4.2 billion yuan — about $619 million — by floating at least a 10% stake, which implies a total valuation near 42 billion yuan, or $6.2 billion. Those are ordinary IPO figures. The figures underneath them are not, because they are black.

For fiscal 2025, Unitree reported revenue of 1.699 billion yuan — roughly $250 million — and a net profit of 278 million yuan, about $41 million. A humanoid company that makes money is close to a category error given how the rest of the field is financed. And the mix is shifting fast in exactly the direction the market cares about: humanoid robots grew from 27.6% of revenue across all of 2024 to 51.5% in the first nine months of 2025. The company shipped more than 5,500 humanoid units in 2025. This is not a prototype shop with a demo reel. It is a manufacturer with a shipping line and a P&L.

Why the profit exists — and where it's fragile

The uncomfortable, clarifying detail is how Unitree makes money, and it is not by selling many $16,000 humanoids to buyers who deploy them at scale. The profitability leans heavily on the company's actuators and its lower-cost quadruped robots — the components and the dog-shaped machines it has sold for years into research labs, universities, and increasingly other robotics companies. Unitree is, to a meaningful degree, profitable as a parts and platform supplier to the humanoid boom rather than as a seller of finished humanoids doing useful work.

That is a genuine strength and a genuine question mark at once. Selling the picks and shovels is the durable position in any gold rush, and Unitree's actuators are a real competitive asset in an industry where the joint is the hard part. But it also means the humanoid revenue that is growing fastest is not yet the profitable revenue, and the valuation is being underwritten against a transition — from selling the parts to selling the robots — that has not been completed. Investors buying the "first humanoid stock" are buying a components business with a humanoid narrative bolted on top, and the two are not yet the same company.

The speed record and what it signals

The listing itself set a record: approval took 104 days from acceptance to registration, the fastest ever under the STAR Market's pre-review mechanism. That is not an accident of paperwork. It is a policy signal. Beijing has designated embodied AI a strategic priority, and fast-tracking the sector's flagship onto a domestic exchange is of a piece with a broader push to build a public-market capital base for Chinese robotics rather than see the best companies chase foreign listings. The IPO is as much industrial policy as it is a financing event.

That framing cuts both ways for outside observers. It means Unitree will list with the wind of state priority at its back — cheap capital, favorable review, a captive investor base primed to treat it as the sector proxy. It also means the stock will trade as a political object as much as an operating business, exposed to the same national-security scrutiny abroad — over data, over dual-use, over China's intelligence laws — that has shadowed every Chinese hardware champion before it. The company that is easiest to buy on the STAR Market may be the hardest to deploy in a Western industrial facility.

The signal for the rest of the field

Strip away the geopolitics and the accounting nuance, and Unitree's debut still matters for a simple reason: it puts a price on humanoids that is set by a public market rather than a private round. Until now, every humanoid valuation has been a negotiated fiction between a founder and a venture fund, marked to a story. A liquid, publicly traded pure-play forces a daily, adversarial answer to the question the whole industry has deferred — what is a robot company that ships worth, when you can no longer defer the payoff a decade and call it upside?

Unitree gets to that answer first not because its robots are the most advanced, but because it is the one that already turned a profit — even if the profit comes from the parts more than the machines. The rest of the field will be watching the tape, because the number the market assigns to the profitable outlier is the number every unprofitable competitor will be measured against next.