Where does the money in humanoid robotics flow first? The answer might be counterintuitive.

On September 5, Wuxi QuanZhiBo announced the completion of Series B funding of several hundred million yuan, led by Yida Capital and Guotai Haitong Kaiyuan, with Xizheng Equity participating. All three existing shareholders — Guanghe Ventures, Shouchang Holdings, and Wuxi Xintou — chose to continue investing.

That's not the most astonishing number about this company. The most astonishing number is: 8 funding rounds in less than two years.

Last July it was still raising its angel round. This July it finished an A+++ round (led by GL Ventures, with ZhiYuan and LingXin QiaoShou participating). September it's already at Series B. Averaging one round every three months. The pace is so fast they're running out of letters, and plus signs.

What QuanZhiBo makes is integrated robot joint modules. Not the sexiest complete robot. Not the most imaginative large model. It's the "joints" inside a robot's body — the things that make a robot's arm turn, leg bend, fingers grasp.

Why Are Joints Exploding First?

The humanoid robotics industry chain roughly breaks down into three layers: top layer is complete robot companies, middle layer is core components (joints, motors, reducers, sensors), bottom layer is upstream raw materials and chips.

Everyone always pays most attention to the top layer — who released a new robot, whose robot can dance again, whose valuation went up again. But the ones making real money and scaling first are usually in the middle layer.

Why? The reason is simple: complete robot companies are still in the demo phase, joint companies are already selling product.

Across the entire humanoid robot industry, everyone's saying "mass production next year" and "100,000 units the year after" — but how many actual bulk orders are there? Not many. Most complete robot companies are still burning money on R&D, demos, and storytelling. But joints are different — regardless of which company's robot eventually sells, as long as robots get built, they have to buy joints.

It's just like the gold rush: the ones who really made big money weren't the gold prospectors, they were the ones selling jeans. In this humanoid robot gold rush, joints are the jeans.

QuanZhiBo's technology roadmap also tells a story. They develop in-house all three major technology routes — "planetary + harmonic + cycloidal" — covering joint products for different load levels. What that means is whether you're building light-load robots or heavy-load robots, whether you use planetary reducers or harmonic reducers, they can supply you. It's like opening a joint supermarket — whatever you need, they've got it.

Mass Production Ramp: The Dividing Line Between "Can Make" and "Can Make a Million"

The money from this round goes mainly to three places: million-unit production capacity ramp, next-gen high torque density joint R&D, and global customer service.

The first keyword is "million-unit production capacity." This word matters. Because there are quite a few companies making robot joints, but not many can do million-unit capacity. Even fewer can do it with consistent supply. There's an old joke in robotics: you see a hundred companies at demo time, but only three to five left at mass production time. That's exactly what this is about — making one or two in a lab is fine. Make ten thousand and consistency breaks down. Make a hundred thousand and yield collapses. Make a million? Most companies don't even dare think about it.

When QuanZhiBo dares say "million-unit production capacity ramp," it means they've passed the "can we make it?" stage and entered the "can we make a million of them?" stage. That's the critical step for a component company from "startup" to "serious supplier."

The second keyword is "high torque density." Torque density is how much torque you can output per unit of weight. This parameter directly determines how heavy a robot can lift, how fast it can walk, how much it can carry. For humanoid robots to truly enter factories and homes, torque density has to go up another level. Current joints are either too heavy, don't have enough torque, or too expensive. Next-gen high torque density joints is what every humanoid robot company is waiting for.

The third keyword is "global customer service." That means QuanZhiBo's ambition isn't just the domestic market. It wants to be a joint supplier in the global supply chain. If you can break into the global humanoid robotics supply chain, the ceiling is completely different.

Everyone's watching the top-layer complete robot companies in the humanoid robot story, everyone's waiting for that "killer app" to appear. But until then, money flows first to the middle layer — to those component companies that can actually ship product, make money, and secure their position. QuanZhiBo's 8 rounds in 2 years isn't a coincidence. It's a signal — in robotics, the ones who make money first are always the shovel sellers.

See you tomorrow.