Unitree Robotics: How a Chinese Underdog Became the King of Robots

By Sam

I first saw a Unitree robot on YouTube at 2 AM.

I couldn't sleep. Scrolling. You know how it goes. Then I saw this thing – a robotic dog that looked like something from a sci-fi movie. It was running. Jumping. Doing backflips. Carrying a person. I thought it was CGI. I really did.

It wasn't.

That dog was real. It's called the Unitree B2. And it costs less than a Tesla.

I fell down the rabbit hole that night. And what I found scared me a little. Not because robots are scary. Because China is moving so fast, and America isn't paying attention.

Let me explain.

Who Is Unitree Robotics?

Unitree Robotics is a Chinese company founded in 2016 by a guy named Wang Xingxing. He was a college dropout. Not a PhD from MIT. Not a former Google engineer. Just a kid who loved robots and refused to stop building.

Before Unitree, Wang built a small quadruped robot in his dorm room. People laughed. He didn't care. He kept going.

Fast forward to 2026. Unitree is now the undisputed king of quadruped robots in China. Their dog robots – the Go2, B2, and AlienGo – are everywhere. Research labs. Construction sites. Police departments. Even people's homes.

And now they're making humanoids. The Unitree H1. A robot that walks, runs, jumps, and balances like a human. It's not perfect. But it's improving every month.

The craziest part? Their robots are cheaper than Boston Dynamics. Sometimes half the price. Sometimes less.

How?

Unitree

How Unitree Became Number 1 in China

Three reasons. And they're not what you think.

1. They focused on one thing only: making robots affordable.

Boston Dynamics spent decades perfecting their robots. They made them incredible. But also incredibly expensive. A Spot robot costs $75,000. That's a luxury car.

Unitree looked at that and said: "What if we make one for $10,000?"

They stripped away everything that wasn't necessary. Used cheaper motors. Simplified the design. Manufactured at massive scale in China's supply chain. The result? A robot that's 80% as good for 20% of the price.

That math works for customers. Labs buy more. Police buy more. Developers buy more. More customers mean more data. More data mean better robots. Better robots mean more customers.

It's a flywheel. And it's spinning fast.

2. They opened up their software to everyone.

Boston Dynamics kept their software locked down. You bought the robot, but you couldn't easily modify it. You needed their permission. Their tools. Their developers.

Unitree did the opposite. They released software development kits. They published APIs. They encouraged developers to hack their robots, build new features, share code.

A community grew around Unitree. Students in Shanghai. Engineers in Berlin. Hobbyists in Texas. All of them building on Unitree's platform. For free.

That's how you win. Not by building everything yourself. By letting the world build for you.

3. They got help from the Chinese government.

I'm not naive. China wants to lead in robotics. It's a national priority. Unitree received grants, low-interest loans, and government contracts. The Chinese military is interested. The police are interested. Public safety departments are buying robots for search and rescue, inspection, and surveillance.

That doesn't make Unitree bad. It makes them smart. If your government wants to help you win, you say yes.

Does Unitree Depend on NVIDIA for Chips?

Yes. Mostly.

Unitree uses NVIDIA chips for their robots. The Jetson series. The Orin. These are powerful AI chips that handle all the computing – vision, balance, decision-making.

But here's where it gets complicated.

China is under US sanctions. NVIDIA cannot sell their most advanced chips to Chinese companies. So Unitree works with what they can get. They use slightly older NVIDIA chips. Or they use domestic Chinese chips from companies like Horizon Robotics and Cambricon.

Unitree is also investing in their own chip research. Not because they want to. Because they have to. If the US cuts off chips tomorrow, Unitree needs a backup.

So the answer is: mostly NVIDIA, but diversifying fast.

Can America Compete with Unitree?

Yes. But not easily.

America has Boston Dynamics. They make incredible robots. The Spot. The Atlas. Stretch. Their technology is probably still ahead of Unitree. More refined. More capable.

But Boston Dynamics has two problems.

First, they're expensive. A Spot is $75,000. A Unitree B2 is $10,000. For most customers, that difference matters.

Second, Boston Dynamics was bought by Hyundai. They're now a Korean-owned company. Still American in spirit. But not fully.

Other American robotics companies exist. Ghost Robotics makes quadrupeds for the military. Agility Robotics makes humanoids for warehouses. Tesla is working on Optimus.

But no single American company has captured the public imagination like Unitree. No one has made robots feel accessible, affordable, and fun.

Can America catch up? Yes. Will they? That depends on whether investors and the government decide it's a priority.

Will Unitree Build Robots for the Chinese Army?

I think they already have.

Not Terminator-style killing machines. Not yet. But logistics robots. Reconnaissance robots. Robots that carry supplies, inspect dangerous areas, or survey battlefields.

The Chinese military is modernizing fast. They see robots as a way to reduce casualties and extend their reach. Unitree's robots are perfect for that.

Does that make Unitree evil? No. Every robotics company faces this question. Boston Dynamics sold robots to American police. Ghost Robotics works with the US military. It's not unique to China.

But we should pay attention. When robots go to war, the rules change.

Is It Good for Humanity? Or Will There Be Robot Wars?

This is the question that keeps me up at night.

I love robots. I think they're incredible. They can search for survivors in collapsed buildings. Inspect nuclear reactors. Deliver medicine to remote villages. Assist elderly people who live alone.

That's the good.

But the same robots can be weapons. Attach a gun to a robot dog. Arm a humanoid with explosives. Send a swarm of drones to attack a city.

That's the bad.

And the ugly? When we stop seeing robots as tools and start seeing them as soldiers. When we remove humans from the decision to kill. When we let algorithms decide who lives and who dies.

I don't think we're there yet. Unitree isn't building killer robots. Not openly.

But the technology is moving fast. Faster than the laws. Faster than ethics. Faster than our ability to have this conversation.

What I Think About the Future

I think robots will be everywhere in 10 years.

Not just in factories. In homes. Hospitals. Schools. Police stations. Battlefields.

I think China will lead in hardware – motors, batteries, bodies. America will lead in software – AI, decision-making, ethics. That's not bad. That's specialization.

I think we need rules. International rules. Treaties that ban autonomous weapons. Laws that require human oversight. Standards that prevent robots from being used for torture or repression.

Without rules, we get an arms race. China builds better robots. America builds better robots. Russia builds better robots. No one wins. Everyone spends trillions. And the first time a robot kills a civilian by mistake, public trust evaporates.

I'm not anti-robot. I'm anti-stupid.

The Human Truth

Here's what I keep coming back to.

Robots don't kill people. People kill people. Robots are just tools. Like guns. Like drones. Like knives.

The question isn't "are robots dangerous?" The question is "are we responsible?"

Can we build robots that serve humanity instead of destroying it? Can we write laws that keep up with technology? Can we resist the temptation to automate violence?

I don't know. I hope so.

But I'm not betting on it.

Final Thoughts

Unitree Robotics is an incredible company. They took a technology that was expensive and fragile and made it affordable and durable. They built a community. They captured the world's attention.

They are not evil. They are not saints. They are a company trying to win.

America can compete. But we have to decide that it matters. We have to fund research. Support startups. Attract talent. Write smart laws.

And we have to have the hard conversations about robots and war and ethics. Not after something terrible happens. Before.

I'll be watching. I'll be writing. And I'll be hoping that the future is better than my 2 AM YouTube nightmares.

– Sam

P.S. I still want a robot dog. Not for anything useful. Just to walk around the neighborhood and confuse my neighbors. That's the dream.

P.P.S. If you work in robotics – thank you. Build something that helps people. Please.

End of article.