How Did Unitree Become the World's First Public Humanoid Robot Company?

On July 7, 2026, something remarkable happened in the global robotics industry: a Chinese company called Unitree Robotics received regulatory approval to list on Shanghai's STAR Market, becoming the world's first publicly traded humanoid robot company. The approval came just 104 days after Unitree filed its application—the fastest review in STAR Market history. The company plans to raise 4.2 billion yuan ($620 million) to expand production and accelerate embodied AI research.

To understand why this matters, you need to understand what Unitree has accomplished. In an industry where Boston Dynamics has spent 30 years developing impressive but wildly expensive robots that almost nobody can afford, Unitree took a completely different approach. It decided to make robots that people can actually buy—starting at $5,000 instead of $500,000.

That bet is paying off. Unitree is already the global leader in humanoid robot shipments. Its 2026 target: 20,000 units. For context, Boston Dynamics has built perhaps a few hundred Atlas robots in its entire history. Tesla's Optimus program is still in development and hasn't shipped meaningful volume.

How did a startup from Hangzhou, founded in 2016 by a then-22-year-old engineer named Wang Xingxing, go from making four-legged robots in a garage to leading the most anticipated technology category of the next decade? Let's break it down.

From Quadrupeds to Humanoids: The Unitree Origin Story

Unitree didn't start with humanoid robots. It started with dogs—robot dogs, that is. Wang Xingxing, a robotics enthusiast from Zhejiang Province, founded Unitree in 2016 with a simple but radical idea: make legged robots affordable.

The Boston Dynamics Problem

For decades, Boston Dynamics defined what legged robots could do. Their videos of BigDog, Atlas, and Spot went viral, demonstrating capabilities nobody else could match. But there was a problem: these robots cost hundreds of thousands of dollars each, and almost nobody could buy one. Boston Dynamics was essentially a research project masquerading as a company, passed between owners like Google, SoftBank, and Hyundai.

Wang Xingxing looked at this and saw an opportunity. What if you didn't need the absolute best performance? What if you built robots that were good enough—and made them cheap enough that universities, research labs, and even hobbyists could buy them?

The Laikago Era

Unitree's first significant product was the Laikago, a quadruped robot released in 2018. Named after the famous Soviet space dog, Laikago wasn't as capable as Boston Dynamics' Spot—but it cost a fraction of the price. At around $45,000, it was still expensive, but it was in a completely different league from Spot's $74,500 price tag (which came years later).

Unitree followed up with the A1, then the Go1, then the Go2—each generation getting cheaper and more capable. The Go2, released in 2023, started at just $1,600 for the base model. Suddenly, anyone who wanted a quadruped robot could afford one.

This relentless drive toward affordability wasn't just a marketing strategy—it was built into Unitree's DNA. The company designed and manufactured almost everything in-house: motors, controllers, sensors, even the structural parts. By controlling the entire supply chain and leveraging China's manufacturing ecosystem, Unitree could drive costs down faster than anyone else.

The Humanoid Pivot: From Go2 to H1

By 2023, Unitree was the world's largest seller of legged robots. But the company had bigger ambitions. In August 2023, it unveiled the Unitree H1, its first full-size humanoid robot. The price? Around $90,000—an order of magnitude cheaper than any comparable humanoid robot on the market.

The H1 wasn't perfect. It lacked hands initially (Unitree shipped them as an optional upgrade). Its movement wasn't as fluid as Boston Dynamics' Atlas. But at $90,000, it was accessible. Universities bought them for research. Companies bought them for prototyping. Even wealthy hobbyists bought them just to experiment.

The Product Lineup Expands

Unitree didn't stop at the H1. Over the next two years, the company released a dizzying array of humanoid robots:

• H2 (2024): The next-generation full-size humanoid, standing 1.82m tall and weighing 70kg. Improved balance, better sensors, and an upgraded AI brain. Shown doing parkour and martial arts moves in viral videos.
• G1 (2025): A smaller, more affordable humanoid at just 80cm tall and 12kg. Priced from 19,999 yuan (~$2,800). Designed as a development platform for AI researchers and hobbyists.
• R1 (2025): A "smart companion" robot weighing about 25kg, with integrated voice and vision AI. Priced from 39,900 yuan (~$5,500).
• Unitree Dual-Arm Humanoid (April 2026): A new lineup of upper-body humanoid robots focused on manipulation tasks. Four models available (R1-A5, R1-A7, R1-A5-D, R1-A7-D), starting at just 26,900 yuan (~$3,700).

Every few months, Unitree releases a new robot or a new version. This pace of iteration is almost unheard of in robotics, where companies typically spend years on each generation. Unitree operates more like a consumer electronics company than a traditional robotics firm.

The $620M IPO: Record Speed, Record Ambition

The IPO announcement in July 2026 shocked the industry—not because Unitree was going public, but because of how fast it happened.

104 Days: The Fastest STAR Market Review Ever

Unitree filed its IPO application on March 20, 2026. It passed its listing hearing in 73 days. And on July 7, it received final registration approval—just 104 days from application to green light. For comparison, most STAR Market IPOs take 6-12 months.

The speed sent a clear signal: the Chinese government considers humanoid robotics a strategically important industry, and Unitree is its champion. The fast-track approval was both practical (Unitree's financials and technology are solid) and political (China wants to lead the embodied AI revolution).

What the Money Will Be Used For

Unitree plans to raise 4.202 billion yuan ($619.8 million). Here's where it's going:

• Expanded production: Building new manufacturing facilities to scale humanoid robot production from thousands to tens of thousands of units per year.
• Embodied AI R&D: Investing heavily in the artificial intelligence that will make humanoid robots truly useful—not just impressive demo machines.
• New products: Developing next-generation humanoids with better manipulation, longer battery life, and more advanced AI capabilities.

The 2026 shipment target of 20,000 units sounds conservative given the pace of orders. If Unitree hits this target, it will have more humanoid robots in the wild than every other company combined.

Commercial Use Cases: From Airports to Factories

The biggest question about humanoid robots has always been: what are they actually for? Unitree is starting to answer that question by deploying robots in real commercial settings.

Tokyo Haneda Airport Deployment

One of the most significant commercial wins for Unitree came in early 2026, when its G1 robots entered trial operations at Tokyo's Haneda Airport—one of the busiest airports in the world. The robots work alongside Japan Airlines ground handling staff, performing tasks like:

• Baggage loading and unloading
• Cargo transport
• Conveyor belt coordination

The trial, scheduled to run through 2028, marks the first commercial application of a Chinese humanoid robot inside a major global aviation hub. It's also significant because Japan—home to robotics giants like Honda, SoftBank, and Toyota—is importing Chinese humanoid robots. That would have been unthinkable five years ago.

Industrial and Logistics Applications

Beyond airports, Unitree robots are being tested in a range of commercial scenarios:

• Manufacturing: Working alongside human workers on assembly lines, performing repetitive tasks like parts handling and quality inspection.
• Logistics: Moving packages in warehouses and distribution centers, navigating around obstacles and working alongside human workers.
• Public services: Greeting visitors in government buildings and shopping malls, providing information and directions.
• Research: Used by hundreds of universities worldwide as a platform for AI and robotics research.

These early deployments aren't about replacing human workers—at least not yet. They're about finding the right tasks where humanoid robots can add value reliably. The fact that Unitree's robots are affordable enough that companies can afford to experiment with them is accelerating this discovery process.

Unitree vs Boston Dynamics: Two Different Philosophies

Comparing Unitree and Boston Dynamics reveals two completely different approaches to building a robotics company. Neither is "right" or "wrong"—they're optimizing for different things.

"Boston Dynamics builds the best robots in the world. Unitree builds the most available robots in the world. And in the long run, availability might be more important than peak performance." — Robotics Industry Analyst

The advantage of Unitree's approach is data. Every robot that ships generates real-world data about how robots move, fall, succeed, and fail. With thousands of robots in the wild, Unitree is collecting more data about legged locomotion than any other company on Earth. That data feeds back into better algorithms, which leads to better robots, which leads to more sales. It's a flywheel that Boston Dynamics, with its low-volume approach, can't easily match.

The Competition: UBTECH, Tesla, Xiaomi, and More

Unitree isn't the only player in the humanoid robot space. The competition is heating up fast, especially in China.

UBTECH: The Old Guard with a New Trick

UBTECH is the elder statesman of Chinese humanoid robots, founded way back in 2012. For years, it was known for making educational robots and the Walker humanoid series. But in July 2026, UBTECH surprised everyone by launching its "Utopia U1" series of hyper-realistic humanoid robots—human companions that cross the uncanny valley with 1:1 high-fidelity reproduction of skin pores and subcutaneous blood vessels.

The top-tier model costs nearly 1 million yuan (~$140,000), but it secured over 13,000 pre-orders from high-net-worth customers in just a few days. UBTECH is taking a completely different path from Unitree—luxury companions instead of industrial workers—but both companies are proving there's real demand for humanoid robots.

Tesla Optimus: The Wildcard

Tesla's Optimus program is the biggest unknown in humanoid robotics. Elon Musk has promised millions of robots at low prices, but the company has yet to demonstrate a robot that can reliably perform useful work. Tesla's advantages—AI expertise, manufacturing scale, brand recognition—are significant. But so is its challenge: building a humanoid robot that's actually useful is harder than almost anyone anticipated.

Unitree's head start—with actual shipping products and real-world data—could be decisive. Or Tesla could come from behind with better AI and manufacturing prowess. Nobody knows yet.

Xiaomi and the Rest

Xiaomi has also shown off its CyberOne humanoid robot, and other Chinese tech companies are rumored to have humanoid programs. The ecosystem of component suppliers, AI companies, and manufacturing partners in China means that new humanoid robot companies can launch and scale faster than anywhere else in the world.

According to a 2026 industry report, China now has over 140 humanoid robot manufacturers, with total shipments reaching 14,400 units in 2025. Unitree alone accounts for a significant share of those shipments.

The Spring Festival Gala Moment: Viral Marketing Mastery

If you want to understand how Unitree captured the public's imagination in China, look no further than the 2026 CCTV Spring Festival Gala—China's most-watched television program, with hundreds of millions of viewers.

Unitree made its third consecutive appearance at the Gala, this time bringing its G1 and H2 humanoid robots for a stunning "Cyber Real Kung Fu" performance. The robots demonstrated feats that would be impressive for human acrobats:

• World's first continuous freestyle table-vaulting parkour by humanoid robots
• Aerial flips with maximum height exceeding 3 meters
• Continuous single-leg flips and two-step wall-assisted backflips
• An Airflare grand spin of seven-and-a-half rotations
• High-speed cluster movement with up to 4 m/s speed for dozens of robots

The performance went viral across Chinese social media. For most Chinese viewers, it was the first time they'd seen humanoid robots doing things that looked genuinely useful (and cool). The Gala appearance cemented Unitree's position as the face of China's humanoid robot industry.

This isn't just marketing fluff. The engineering behind those stunts—balance, coordination, agility, cluster synchronization—directly translates to more useful real-world robots. If a robot can do a backflip, it can definitely walk across an uneven factory floor.

The Embodied AI Opportunity

The real long-term value of humanoid robots isn't in the hardware—it's in the software. This is where "embodied AI" comes in: artificial intelligence that exists in a physical body and interacts with the real world.

For years, AI has existed only in the digital world—chatbots, image generators, recommendation algorithms. It's impressive, but it can't do anything physical. Embodied AI changes that by putting AI in robots that can move, touch, manipulate, and navigate the real world.

Unitree's position as the volume leader in humanoid robots gives it a front-row seat to this revolution. Every robot that ships is a data-gathering platform. Every time a robot walks, falls, or successfully completes a task, Unitree learns something. Over time, this data advantage could be more important than any single hardware innovation.

The company's IPO prospectus emphasizes "strengthening embodied artificial intelligence research and development" as a key use of funds. This isn't an afterthought—it's the core of Unitree's long-term strategy. The company that builds the best physical platform AND the best AI brain will win the humanoid robot race. Right now, Unitree is the only company with both a real hardware platform and significant real-world data.

Challenges and Risks

For all its successes, Unitree faces significant challenges ahead:

Commercialization Risk

The biggest question about humanoid robots is whether they can do useful work at a cost that makes economic sense. Right now, most humanoid robots are used for research or novelty. Unitree needs to prove that its robots can deliver real ROI in industrial, logistics, or service settings.

AI Capability Gap

Walking and doing backflips is impressive. But what people really want from humanoid robots is intelligence—the ability to understand complex instructions, adapt to new environments, and perform useful tasks without extensive programming. Unitree's AI capabilities are improving, but they're not at the level of companies like Google DeepMind or Tesla.

Competition Intensifying

The humanoid robot space is getting crowded fast. Tesla, Google, Amazon, Xiaomi, UBTECH, Agility Robotics, Figure AI—everyone is getting into the game. Unitree's first-mover advantage is real, but it could be eroded by competitors with deeper pockets or better AI.

Quality at Scale

Making 1,000 robots is hard. Making 20,000 robots is harder. Making millions of robots that work reliably? Nobody has done that yet. Unitree's manufacturing capabilities will be tested as it scales up production.

Conclusion: The First, But Not the Last

Unitree's IPO is a milestone for the humanoid robot industry. It's the first public company focused primarily on humanoid robots, and its rapid approval signals both technological maturity and strategic importance.

But Unitree won't be the last. The embodied AI revolution is just beginning, and the companies that win will be the ones that combine good hardware with good AI at a price people can afford. Right now, Unitree is further along this path than anyone else.

The story of Unitree is also the story of modern Chinese tech innovation: start with a category defined by Western companies, figure out how to make it affordable, iterate faster than anyone else, and then use that volume advantage to build a data and manufacturing moat. We've seen this pattern with drones (DJI), EVs (BYD), and batteries (CATL). Now we're seeing it with humanoid robots (Unitree).

Will Unitree become the "Tesla of humanoid robots"? It's too early to say. But one thing is clear: the humanoid robot era is beginning, and a company from Hangzhou is leading the charge. The IPO is just the starting line.