Unitree Raises $620M in a Shanghai Humanoid IPO 2026

A nine-year-old company that started by selling a $1,500 robot dog is about to become the first pure-play humanoid maker to ring the bell on a major exchange. Unitree Robotics has cleared the listing review for an initial public offering on Shanghai's STAR Market, seeking roughly $620 million at a valuation that could mark the moment humanoid robots stopped being a demo and started being a public-market asset class. The numbers underneath the filing are stranger and more revealing than the headline, because Unitree is going public exactly as its growth is cooling.

What Actually Happened

Unitree filed to raise about 4.2 billion yuan, roughly $620 million, on the Shanghai Stock Exchange's STAR Market, the mainland board built for hard-technology companies. The exchange accepted the application on March 20, 2026 and advanced it to the listing review committee in just over two months, with the committee meeting in early June to examine the offering. That pace makes it one of the fastest IPO reviews on the STAR board this year, a signal that Beijing wants a humanoid champion public and wants it quickly. At a public float of no less than 10 percent, the raise implies a baseline valuation near 42 billion yuan, about $6.2 billion.

The financial trajectory behind the filing is steep. Revenue climbed from 159 million yuan in 2023, about $23.5 million, to 1.70 billion yuan in 2025, roughly $250.6 million, a compound annual growth rate near 227 percent. Humanoid robots became the core revenue pillar, with annual shipments topping 5,500 units, the most of any company worldwide, contributing 868 million yuan, about $128 million, or 51.78 percent of total revenue. The robot-dog business that made Unitree famous has been quietly overtaken by the bipedal machines that now define its valuation.

Founder Wang Xingxing remains firmly in control. Born in Ningbo and raised in Zhejiang, Wang started Unitree in August 2016 and holds about 23.82 percent of shares directly, a roughly 33 percent economic interest once indirect stakes are counted, and 68.78 percent of voting rights before the IPO. That concentration matters: Wang can take the company public, raise growth capital, and still steer it like a founder-controlled startup. The offering also arrives weeks after Nvidia chose Unitree's nearly six-foot H2 humanoid, paired with Nvidia's Jetson Thor and Blackwell hardware, as the first robotics system it would sell to research labs from Stanford to ETH Zurich.

Why This Matters More Than People Think

For two years the humanoid story has run on staged videos: robots folding laundry, doing backflips, walking across a stage at a keynote. None of that tells you whether the category is a business. Unitree's prospectus does. It shows a company shipping thousands of bipedal units a year and booking real revenue against them, which converts the humanoid narrative from speculation into an income statement that public investors can underwrite. The first humanoid IPO sets a reference price for an entire sector that, until now, has been valued entirely on pitch decks and private rounds.

The deeper shift is that the center of gravity in humanoid hardware has moved to China, and this listing makes it official. Unitree ships more humanoid units than any rival on earth, undercuts Western competitors on price by a wide margin, and is now tapping public capital markets at home to fund the next phase. Where American humanoid leaders raise enormous private rounds and stay private, Unitree is choosing the discipline and the liquidity of a public listing. That choice gives it a currency, its own stock, to acquire talent and suppliers, and a public valuation that anchors every future financing.

There is a manufacturing logic here that software-first observers tend to miss. Humanoids are not won in the lab; they are won on the assembly line, in the supply chain for actuators, harmonic reducers, and battery cells, and in the cost curve that comes from building at volume. Unitree sits inside the Chinese industrial base that already dominates drones, electric vehicles, and batteries, which means it can drive unit costs down faster than any competitor without that ecosystem. A public listing pours capital into exactly the part of the problem, scaled production, where China holds its strongest structural advantage.

There is a financing-discipline angle as well. A public listing subjects Unitree to quarterly reporting, audited margins, and the unforgiving feedback of a daily share price, which is the opposite of the patient, hype-tolerant private capital funding its Western rivals. That discipline can sharpen a company, forcing it to prove unit economics rather than narrate them, but it also exposes every stumble in real time. By choosing the public route now, Wang is betting that Unitree benefits more from the credibility and liquidity of a listing than it loses from the scrutiny, a wager that only makes sense if he believes the manufacturing numbers will keep validating the story quarter after quarter.

The Competitive Landscape

Unitree's most visible Western rival is Figure, which exceeded $1 billion in a Series C at a reported $39 billion post-money valuation, a figure more than six times Unitree's implied IPO value despite Figure shipping a fraction of the units. Tesla's Optimus program promises mass production but has yet to deliver external volume at scale. Agility Robotics, Boston Dynamics under Hyundai, and a wave of Chinese challengers like UBTech, AgiBot, and Galbot round out a field that has raised tens of billions in aggregate. Unitree's pitch is the opposite of the American one: not the most capable demo, but the most units actually sold.

The valuation gap is the story within the story. Figure is privately valued at roughly $39 billion on the promise of future deployment, while Unitree is going public at about $6.2 billion on the basis of actual shipments and revenue. The market is pricing the American narrative of breakthrough capability against the Chinese reality of manufacturing throughput. Industry observers in China argue Unitree could be worth more than 100 billion yuan, over $14 billion, once public, which would still leave it a third of Figure's private mark while selling far more hardware. Public markets are about to adjudicate which model deserves the premium.

The historical parallel is the early electric-vehicle market. A decade ago, Tesla carried a valuation built on vision and software while Chinese manufacturers like BYD were dismissed as low-margin hardware shops. BYD then out-produced everyone, drove costs through the floor, and became the volume leader of the global EV transition. Unitree is running the BYD playbook for humanoids: win on manufacturing, price, and volume rather than on the flashiest demo. Whether that playbook repeats will determine whether the humanoid era is led from Silicon Valley or from Hangzhou and Shenzhen.

The component layer is where this contest is quietly decided. A humanoid is a stack of actuators, harmonic and planetary reducers, force sensors, battery cells, and compute, and China already manufactures most of those inputs domestically at volume. Western humanoid makers often import the same parts from Chinese suppliers, which means their cost structure is partly hostage to the very ecosystem Unitree sits inside. The risk for Figure, Tesla, and Agility is that they win the software and perception race yet still pay more for every joint and motor than a Hangzhou rival that buys those components down the street. Public capital lets Unitree vertically integrate that supply chain even further.

Hidden Insight: Unitree Is Going Public Precisely as the Growth Cools

The uncomfortable detail buried in the filing is the deceleration. After years of triple-digit expansion, Unitree's first-quarter 2026 revenue growth slowed to 68 percent from 332 percent a year earlier, and adjusted net profit fell 52.55 percent year over year. A company that grew nearly 227 percent annually is suddenly growing a fraction of that, and its profit halved on the eve of the offering. The bear case is straightforward: Unitree may be tapping public markets at the exact moment its hyper-growth phase ends, locking in a valuation that prices the old curve rather than the new one.

That timing is not necessarily cynical, but it is informative. The profit squeeze almost certainly reflects heavy reinvestment into humanoid research, where margins are thin and the bill of materials is still expensive, even as the legacy quadruped business matures. In other words, Unitree is trading near-term profit for a bet on the humanoid future, and it wants public capital to fund that bet before the market fully reprices the slowdown. A founder with 68.78 percent voting control can make that trade unilaterally, which is both the strength and the risk of a company this concentrated.

The most non-obvious read is that revenue concentration cuts two ways. Humanoids now generate more than half of Unitree's sales, which is the basis of the bull case, but most of those units sell to research labs, universities, and developers rather than to factories running them in production. The G1 deployed at Tokyo's Haneda Airport, where Japan Airlines is trialing it for ground handling through 2028, is the showcase, yet a single airport trial is a pilot, not a fleet. The question public investors must answer is whether Unitree's 5,500 units represent durable industrial demand or a research-and-novelty market that saturates.

This is where the slowdown and the customer mix collide. If the early humanoid buyers were labs and enthusiasts stocking up on the first affordable bipedal platform, that demand is finite and front-loaded, which would explain a sudden growth dip. If instead the deceleration is a temporary pause before factories, logistics firms, and service operators begin buying at scale, the dip is noise before a far larger wave. The IPO forces this question into the open, because public quarterly reporting will reveal, every ninety days, whether humanoid demand is broadening into real industrial use or stalling at the hobbyist and research frontier.

History offers a cautionary template here from the consumer-drone market, which Unitree-adjacent Chinese firms also came to dominate. Drones followed a curve where an early surge of enthusiast and developer demand was mistaken for an infinite market, prices collapsed, weaker players were wiped out, and only the lowest-cost manufacturer, DJI, captured durable profit. Humanoids could rhyme with that arc. If Unitree is the DJI of humanoids, the slowdown is just the shakeout before dominance. If it is one of the many drone makers that did not survive the price collapse, the IPO valuation is a peak. The same manufacturing strength that makes Unitree formidable also invites the brutal price competition that thins the field.

What to Watch Next

In the next 30 days, watch the pricing and the first-day trade. A listing that prices toward the industry's whispered 100 billion yuan figure rather than the baseline 42 billion yuan would signal that public investors are willing to pay private-market multiples for humanoid exposure. Watch the order book and the lock-up terms for Wang and early backers, and watch whether STAR Market regulators impose any conditions tied to the profit decline. The gap between the filing valuation and the trading valuation will be the cleanest public verdict yet on how the market prices humanoids.

Over 90 days, the first earnings report as a public company is the event that matters. Investors will finally see, on a fixed schedule, whether the Q1 2026 deceleration was a blip or a trend, and whether humanoid units are converting from research sales into production deployments. Watch the customer disclosures for any move from pilots like the Haneda trial toward multi-unit factory or logistics contracts. Watch gross margin on humanoids specifically, because the entire bull case rests on Unitree driving the bill of materials down as volume rises.

Over 180 days, the strategic question is whether a public Unitree consolidates the Chinese humanoid field or invites a price war that compresses everyone's margins. With public stock as currency, Unitree could acquire smaller rivals or lock up component suppliers, while competitors like AgiBot and UBTech respond with their own capital raises. Also watch the geopolitical overlay: a Chinese humanoid leader supplying robots and pairing with Nvidia hardware sits squarely in the path of US export policy, and any tightening of controls on advanced robotics chips would land directly on Unitree's roadmap.

Everyone else sells the humanoid future in videos. Unitree is the first to sell it on an income statement, slowdown and all.

---