Just now, another 1 billion yuan in financing: Selling shovels to robots, with a 200% growth.

On April 8th, another huge round of financing was finalized.

Digua Robotics announced the completion of a $150 million (approximately 1 billion yuan) Series B2 financing. It has only been a little over 20 days since the previous Series B1 financing of $120 million. So far, the company has raised a total of $270 million (approximately 1.85 billion yuan) in Series B financing.

In this round of financing, there are both strategic investors such as a retail technology and supply chain giant and Prosperity7, as well as financial investment institutions like Yunfeng Capital and Huakong Fund. Existing shareholders such as Hillhouse Ventures and Wuyuan Capital also participated in the follow - on investment.

Investing 1.85 billion yuan in 30 days, what opportunity are the capitals vying for?

A clue lies in the company's business. It doesn't manufacture complete robot units but provides a capability to make robots smarter.

01 200% Growth

More informative than the financing is its business changes in the past year: In 2025, the shipment volume increased by 180% year - on - year, the number of customers increased by 200% year - on - year, and it has supported the launch of over 100 robot products.

In other words, it's not just making one or two products but supporting a group of robot companies.

Another change is the scale of developers: There are over 100,000 developers globally, covering more than 20 countries and regions in Asia - Pacific, Europe, North America, etc.

In a nutshell, Digua Robotics' business is about "how to make robots smarter".

Specifically, it has three capabilities.

1. Computing power platform (RDK S600): Enables robots to handle complex tasks.

2. Model capabilities.

Allows robots to understand the environment and think about what to do next.

3. System integration: Connects actions, perception, and decision - making.

These are unified into a concept: The embodied intelligent brain base (robot brain). The same set of capabilities can be applied to different robots ("one brain, multiple forms").

02 New Demands

Recently, why are there more such companies? Behind this is actually a change in the robot industry.

In the past, robots were more like mechanical products: competing in structure, hardware, and manufacturing.

For example, Fanuc from Japan, KUKA from Germany, ABB from Switzerland, and Yaskawa Electric from Japan (the "Big Four") have dominated the global industrial robot market for decades, relying on the hardware barriers of precision reducers, servo motors, and control systems.

But now, the situation is changing: A new market is emerging, and the requirements for the "shovel sellers" have changed.

In 2025, the market size of humanoid robots in China was 8.239 billion yuan, and it is expected to exceed 80 billion yuan by 2030. The shipment volume in 2025 was about 5,000 units, and it is expected to reach 50,000 - 100,000 units in 2026.

Robots are becoming more like a "terminal running models". What determines their capabilities is no longer just the hardware, but rather: whether the computing power is sufficient, whether the model is strong, and whether the system can run smoothly.

This is very difficult.

Training an embodied intelligent model requires a large amount of real - world scenario data. The costs of data collection, computing power, and talent are making more and more companies choose to use "ready - made capabilities" instead of building a full - stack system on their own.

It's easier to understand this model if we place it in a larger technological cycle.

In each technological wave, there are two types of companies: one type makes products, and the other type provides basic capabilities.

In the PC era, Intel made chips, Microsoft made operating systems, and IBM and HP made complete machines. In the mobile Internet era, ARM made architectures, Qualcomm made chips, Google made Android, and Xiaomi and OPPO made mobile phones.

In the field of AI, this division of labor has already emerged: NVIDIA provides computing power, OpenAI develops models, and AWS and Azure offer platforms.

Now, a similar structure is taking shape in the robot field.

In China, complete machine manufacturers such as Zhiyuan Robotics, Zhujidongli, and Xingdong Jiyuan are actively raising funds, while underlying platforms like Digua Robotics and Horizon continue to receive capital support.

03 Profit - making Ability

One characteristic of such companies is that they can start making money without waiting for the industry to fully mature, and there are several reasons for this.

Not Betting on Specific Product Forms

Whether it's humanoid robots, quadruped robot dogs, or floor cleaners, lawn mowers, and industrial robotic arms, they all need underlying intelligent capabilities.

For example, Digua Robotics' customers cover consumer - grade products and cutting - edge embodied intelligent robots, with a wide range of product categories and diversified risks.

Can Serve Many Customers

A single system can be reused by multiple robot products. This can be seen from Digua Robotics' data: It serves 100,000 developers and more than 500 startup companies.

Solving the Most Difficult Part

Robot development involves chips, algorithms, systems, and toolchains, with extremely high thresholds. The value of an external solution is directly proportional to the cost of self - building.

The "shovel - selling" business is quite stable, but there are also certain risks, that is, the gold mine must really exist, and the miners must really need shovels.

In other words, its development premise is that robots must truly move towards large - scale production.

If the industry remains in the "demonstration" and "small - scale pilot" stage for a long time, the demand for underlying capabilities will also be limited. In 2025, the shipment volume of humanoid robots was about 5,000 units, and it is still in the process of large - scale production.

In addition, there are two variables:

The Technological Route is Changing Rapidly

From rule - driven to deep learning and then to end - to - end large models, the technological paradigm may change suddenly. Today's mainstream solutions may be eliminated tomorrow.

Competition Pressure

Chip manufacturers (NVIDIA Jetson, Qualcomm RB series), large - model manufacturers (OpenAI, Google), and vehicle manufacturers (Tesla Optimus) may directly enter this field.

Looking back, in each technological wave, the companies that take the lead are often not the most prominent product companies but those that provide underlying capabilities.

The robot industry may also follow this path. It's just that this time, the ore veins are still forming, and the "shovel sellers" have already set off.

This article is from the WeChat official account "Pencil News" (ID: pencilnews), author: Cheng Shide. It is published by 36Kr with authorization.