Stories
The robots are lining up to ring the bell
By Sebastián Ocampo · July 26, 2026 · 6 min read · ES·FR·EN
For a decade, a robot's proof of life was a video. Starting this summer, for China's humanoid companies, it will be a quarterly report. And quarters have no outtakes.
This July's image in Shanghai is not a robot landing a backflip: it is a line of founders heading for the trading floor. On July 3, China's regulator gave final approval for Unitree Robotics to list on the STAR Market after a 73-day review, the fastest on record. The company plans to raise about 4.2 billion yuan, $618 million, by selling at least 10% of its shares, a deal valuing it around 42 billion yuan, nearly $6 billion, with a debut possible before the month ends. Behind it, in the same week, LimX Dynamics closed nearly $200 million in a pre-IPO round valuing it at 15 billion yuan, and Deep Robotics and Leju Robot started their own processes.
A stock prospectus demands what no stage ever does: audited revenue, risks in writing and a price, every quarter, on the distance between demo and delivery.
The man at the head of the line is known in this house. Wang Xingxing founded Unitree in Hangzhou in 2016, controls about a third of the company, and a year and a half ago starred in the night China took its robots dancing: his humanoids dancing on the most-watched television gala on Earth. That night was state marketing; this IPO is something else. A stock prospectus demands what no stage ever does: audited revenue, margins, customer concentration, risks enumerated in writing. And Unitree's numbers tell a story less epic and more solid than its videos: the company has been profitable for years, but what sustains it are quadrupeds and components, not the humanoids that appear at galas.
The sentence that sums up the moment came from LimX founder Will Zhang, speaking to CNBC: "Listing is a must." A must not because the companies are ready, but because the window is open: venture capital has flooded Chinese robotics, valuations have multiplied within months (LimX has raised $400 million in half a year) and every founder knows capital windows close without notice. The precedent everyone cites lives in this house: UBTECH, the first humanoid company to list, joined the Hong Kong exchange in December 2023 and has since used public shares to fund bets like its U1 companion robot.
Here this publication takes out its magnifying glass, and this time the glass works in everyone's favor. Our trade is separating the real robot from the stage robot, and an IPO is the industry's only format that cannot be edited: if Unitree's humanoids generate little revenue, the prospectus will say so; if one client concentrates half the sales, the prospectus will say so; if the promise of household agents is years away, the prospectus will have to say it on the risks page. Every quarter, the market will price the distance between demo and delivery, which is exactly what this publication measures with words. We are about to get an ally with auditors.
For the reader with feet on the ground, the practical translation: public money accelerates factories, and factories lower prices. Unitree already sells the G1 for about $16,000 and the R1 for $5,900, figures unthinkable three years ago, and a scale war funded on the stock market pushes in only one direction. The symmetric risk exists too: if valuations run far ahead of humanoid revenue, the correction will splash the whole sector, honest companies included. Bubbles do not discriminate; prospectuses do.
What to watch has dates and numbers. Unitree's debut day: not the percentage pop, but how much the prospectus reveals about humanoids' real weight in its revenue. The first quarterly results as a listed company. Whether LimX, Deep Robotics and Leju complete their listings or the window closes first. And the number this publication will follow with most interest: the price of the next consumer humanoid. If the bell rings to fund factories, we will all see it on the price tag.