The robot's dream lies in Shenzhen, while the pricing power is in Zhengzhou.
The end of 2025 was more bustling and more brutal than ever before.
In the past couple of days, the atmosphere in the WeChat groups of the robotics circle has become extremely delicate. On the surface, everyone is forwarding the release video of Shenzhen Zhongqing's T800 and praising that amazing mid - air roundhouse kick. But in private dinner gatherings, all practitioners are discussing only one topic, and with deep anxiety: starting at 180,000 yuan.
If this isn't enough to make traditional manufacturers feel suffocated, don't forget there's also Yushu Technology, a well - known price slasher, eyeing the market. Their G1 has slashed the price down to 99,000 yuan.
Thinking about this matter carefully, it's actually quite surreal, and even a bit unsportsmanlike.
Recall that just a year ago, humanoid robots were like precious laboratory darlings worth millions of yuan, exclusive toys of top universities and research institutions. But today, they have become industrial consumables priced the same as a ride - hailing car.
This price drop can't even be described as a discount. It's a halving, or even a devastating price cut.
Who is the driving force behind this? Is it the geeks in Nanshan, Shenzhen? Or the scientists in Haidian, Beijing? No.
If you only focus on Beijing, Shanghai, Guangzhou, and Shenzhen, you won't understand this war. The real answer lies in the seemingly low - key but actually powerful central backbone of China's map.
In the previous two articles, we talked about the speed of the Pearl River Delta and the precision of the Yangtze River Delta. In this article, "The Cost War", we need to turn our attention to Zhengzhou, Wuhan, and Wuhu.
It is this central industrial belt that is using the same ruthless methods used to manufacture iPhones and cars in the past to drive down the price of robots.
Research and development are in Shenzhen, but the steps must be taken towards Zhengzhou
To understand the brutality of this cost war, we first need to do some calculations. Once you figure out these numbers, you'll know how profitable or helpless the robot business used to be.
If you disassembled a mainstream humanoid robot in 2023, you'd find it was like a precision instrument made of gold.
Looking at the bill of materials, you'll see three money - guzzlers: Planetary roller screws, frameless torque motors, and harmonic reducers.
These three components alone account for 60% of the total cost of the robot.
Especially that unassuming screw. In the past, it had to be imported from Switzerland or Germany. It required extremely high precision and had a very low production volume. One screw cost 20,000 yuan. A robot needed more than a dozen of them. Just the cost of the screws was enough to buy a small car.
At that stage, everyone was making prototype machines in office buildings in first - tier cities. Engineers were carefully debugging, the yield rate was extremely low, and the supply chain was as scattered as stars in the sky. At that time, selling a robot for 500,000 yuan? That was like doing charity, as it couldn't even cover the R & D costs.
But what Zhongqing's T800 has done this time really shows an understanding of the situation.
Its press conference was held in the glamorous Shenzhen, with a very futuristic PPT. But its production base was unhesitatingly chosen in Zhengzhou, Henan.
Shortly before the release, Zhongqing announced the establishment of a 200 - mu manufacturing base in Zhengzhou, with a planned annual production of 5,000 units.
Why Zhengzhou? Many people's first reaction is that the land price is cheap. But this is just the surface. Zhengzhou holds a trump card: the world's most mature electronics manufacturing and processing system. There's Foxconn, hundreds of thousands of well - trained industrial workers accustomed to precision assembly, and a well - connected railway logistics network.
The industrial logic behind this is very cold - blooded: In the end, the cost war of robots is a scale war.
In Nanshan, Shenzhen, where every inch of land is precious, it's suitable for innovation and trial production. It's the place to define products. But to mass - produce 180,000 - yuan robots in the tens of millions like iPhones, a move is necessary.
You have to go to the central region, where the land is more expansive, energy is cheaper, and the labor force is more concentrated. Only there, on those roaring assembly lines, can the processing cost of each screw be squeezed to the physical limit. Shenzhen defines the product, but Zhengzhou determines whether you can afford it.
The specialty here isn't duck necks, but the dimensionality - reduction massacre of the automotive supply chain
In addition to geographical advantages, the central industrial belt holds a real trump card, a dragon - slaying sword that can scare all traditional robot companies: the automotive supply chain.
Looking at China's industrial map, you'll find that the central region is not only a transportation hub but also the heart of China's automotive industry.
Wuhan is the headquarters of Dongfeng, Wuhu is the hometown of Chery, and Hefei is a major center for new energy vehicles. When these automotive giants start making robots, it's like a dimensionality - reduction massacre for traditional robot manufacturers.
Looking at the current situation, it's like barbarians knocking on the door: Dongfeng Motor (Wuhan) has announced that its humanoid robots will start working in factories in 2026, directly doing screw - tightening jobs. Chery (Wuhu) has just showcased the Mojia robot, also aiming for mass production.
In the eyes of these central automotive companies, what are robots? Forget about the so - called sacred silicon - based life. In their view, it's just an electric vehicle with two legs.
Since it's a vehicle, build it like a vehicle, and rewrite all the rules. In the past, motors for robots had to be custom - made and were extremely expensive. Now? Just go to the workshop production line, take out a vehicle - grade motor, and modify it. With an annual production volume in the tens of millions, the cost is immediately cut in half.
There's no need to develop new batteries either. The solid - state battery used in Zhongqing's T800 has a battery life of 4 - 5 hours. This is a technology developed by new energy vehicle companies to solve the "range anxiety" problem. Now it's directly used in robots, which is called technology spill - over.
As for the torso, in the past, it had to be milled one by one using CNC machines. Now, following Tesla's example of "integral die - casting", with a single press, the magnesium - aluminum alloy frame is formed, eliminating the need for fine - tuning.
For these people used to shipping products in the tens of millions, reducing the price of a 20,000 - yuan imported screw to 500 - yuan domestic price is just normal operation. They are best at turning luxury goods into industrial products.
When Zhongqing's T800 uses aviation - grade magnesium - aluminum alloy integral die - casting technology, its cold - shining metal shell really looks like Iron Man. But the cost - building logic of this armor is full of a strong automotive flavor. This isn't about building a super - hero worth hundreds of millions. It's clearly about building a two - legged car.
Extreme frugality: Solving rich problems with poor methods
If the transfer of the industrial chain is an open strategy, then the choice of the technical route shows the unique shrewdness and pragmatism of Chinese factories. To drive the price down to 99,000 yuan, Chinese manufacturers have demonstrated amazing calculation abilities, and can even be said to be pushing the limits of frugality.
First, the radar is cut. In the early days, robots had several lidars on their heads to see the way. A good lidar could cost >$1000. But on Yushu's G1, it's hard to see those expensive large radars anymore. Instead, there are cameras that cost just a few dozen yuan each.
The logic is simple: since my AI large - scale model (the "brain") is getting smarter, the requirement for the "eyes" can be lower. Use the free software to make up for the expensive hardware. Even if I'm seeing 2D images, I can use algorithms to visualize the 3D world. This cut reduces the cost by tens of thousands.
Then, the material is changed. In the past, everyone insisted on using carbon fiber for lightness. Carbon fiber is good, but it's a luxury approach. You'd feel heart - broken even with a slight bump. Now? Central factories tell you that engineering plastics (PEEK) are good, and even modified nylon works. As long as the structural design is good and topological optimization is in place, plastic can bear hundreds of pounds, is resistant to falls, and replacing a shell only costs about a hundred yuan.
This practical approach of making do, although it may not sound so high - end and even a bit rustic, is the core secret of how Chinese manufacturing turns all high - tech products into bargain - priced goods. In the business world, being good enough is often more powerful than being the best.
The central region decides the outcome, and robots become household appliances
The prices of 180,000 yuan for the T800 and 99,000 yuan for the G1 mark a turning point in an era: Robots are no longer celestial artworks. They are becoming large - scale household appliances.
Recall that 99,000 yuan is the price of a BYD Qin, and 180,000 yuan is the two - year salary of a highly skilled worker (if they don't eat or drink). When the price of robots approaches or is lower than a worker's annual salary, the math of replacing humans with machines finally makes sense.
In the past, factory owners didn't buy robots because they couldn't afford them, and it was too expensive to repair them when they broke down. Now, Zhongqing tells you that the comprehensive operating cost of these robots is one - third of human labor; Yushu tells you that when it breaks down, you can just replace a joint module for a few hundred yuan. If you were a factory owner, what would you choose?
The end - game of this war will definitely be in the central region.
The situation of this game is clear: the Pearl River Delta is responsible for "giving birth" to robots (innovation, prototype development, product definition), the Yangtze River Delta is responsible for refining robots (control, materials, overcoming difficulties), and the central industrial belt will be responsible for spreading robots across the world.
In the future, robots will be shipped from this central backbone to the world, just like iPhones produced in Zhengzhou and cars produced in Wuhan.
They don't need a soul. They just need to be cheap, durable, and in sufficient supply.
Finally
From the lights of Shenzhen Bay to the precision workshops of the Yangtze River Delta, and finally to the roaring assembly lines in the central region, we can clearly see a cost - reduction roadmap for Chinese robots. The West proposes the concepts (Boston Dynamics, Tesla), while Chinese central factories turn these concepts into bargain - priced industrial products.
When the T800 is being mass - assembled in the Zhengzhou factory and Dongfeng's robots start working in the Wuhan workshop, we know that the era of large - scale commercial use has really arrived.
The body is built, and the price is brought down. Now, there's only one last and most difficult step left: How to install a real soul into this cheap but strong body? After all, if a robot only has a 99,000 - yuan body but a stupid "brain", it's still just a large - scale toy.
In the next article, we'll enter the critical stage of this war - "The Intelligence War". We'll talk about the bubbles and truths behind large - scale models, reinforcement learning, and embodied intelligence.
Note: This article is written based on the publicly available version and industry analysis. Some functional details may be adjusted with product iterations.
This article is from the WeChat official account "Technology Mustn't Be Cold", author: Balang, editor: Fangshan. Republished by 36Kr with permission.