36Kr Exclusive | Former Core Members of DJI Enter the Consumer CNC Market, Securing Nearly 100 Million Yuan in Investment from Meituan, Kunlun Capital, and Miracle Plus

Author | Zhang Ziyi

Editor | Yuan Silai

Yingke learned that the consumer-grade desktop CNC and intelligent digital manufacturing platform "Infimaker" recently announced that it has received multiple rounds of financing in a row, with the financing amount approaching 100 million yuan. The investment institutions include Meituan Strategic Investment, Kunlun Capital, and Miracle Plus. The funds from this round will be mainly used for product R & D, supply chain construction, and high-quality mass production and delivery.

Infimaker was founded in November 2024 by several former core members of DJI and returnee doctors in the robotics field. The consumer-grade CNC (Computer Numerical Control) machines that the team is targeting also represent the last piece of the puzzle in the Maker track.

After consumer-grade 3D printers and consumer-grade laser engravers entered the mass market, consumers have become more aware of creative tools and have higher requirements for ease of use and lower usage thresholds. In the past, CNC products were mainly used in the industrial-grade market, and there are both thresholds and market potential for consumer-grade transformation.

Xie Bowen, the founder and CEO of Infimaker, told Yingke: "Globally, more and more people are beginning to pursue personalized creation and small-scale production. In the United States alone, more than 130 million people consider themselves Makers. However, for individual users and small studios, desktop manufacturing tools that truly have industrial-grade precision, multi - material processing capabilities, and are easy to use are still very scarce."

In the view of Infimaker, the biggest problem with traditional CNC is not just the hardware itself, but that the software and interaction logic still remain in the industrial era.

Past CNC products essentially served industrial manufacturing scenarios and often required users to understand "design for manufacturability" in advance - that is, users had to understand the machine's processing limitations, tool paths, and physical boundaries before drawing.

For ordinary creators, this industrial - era usage paradigm still has a very high threshold.

To lower the usage threshold, Infimaker chose to start with "five - axis linkage". In the CNC processing scenario, the five - axis can reach any position in space, enabling the processing of complex objects such as turbines and human - like wrinkles. This "full spatial reach" ability allows users to avoid frequent manual turning and clamping, greatly increasing the possibility of creation. Especially with the help of AIGC technology, 3D models generated by AI often do not have the "processing logic" of traditional engineering, and only five - axis linkage can transform these imaginative curved surfaces into physical entities with a low threshold.

Moreover, the five - axis linkage function of Infimaker allows its products to have approximately 72.8% more effective processing space than mainstream competitors, enabling one - time processing of larger - sized raw materials and allowing more creativity and parts to be free from the constraints of raw material size.

Xie Bowen believes: "Machines should not be tools that limit creativity. We hope that users have ideas first, and then the algorithms and systems adapt to people, rather than the other way around."

In Xie Bowen's view, past manufacturing tools always required people to understand machines. However, the next - generation creation tools should do the opposite.

Therefore, in addition to lowering the product usage threshold at the hardware level, Infimaker is also trying to redesign the software experience of consumer - grade CNC.

In the traditional numerical control field, learning a CAM software usually takes more than a year, and it is difficult to ensure that usable tool paths can be programmed.

Yingke learned that Infimaker can automatically generate tool paths, perform intelligent recognition, and compensate for errors through its self - developed CAM software. This system not only includes deep - level algorithms such as tool path compensation and vibration compensation but also integrates visual perception functions. With its 6 - position automatic tool changer (ATC) and ruby centering instrument, users only need to put the materials into the enclosed processing chamber, and the entire process of tool setting, centering, and cutting does not require manual intervention.

In order to allow the "desktop factory" to truly enter studios, offices, and even more open spaces, the team has reduced the cutting noise to about 70 dB. The enclosed processing chamber can reduce dust and also reduce damage to the respiratory tract.

"We believe that the end - game is not a single hardware product, but a creation ecosystem formed around personal manufacturing," Xie Bowen said. Infimaker plans to focus on hardware - related revenue in the early stage, including the main unit, accessories, tools, and materials. However, in the long - term plan, more important future revenue will come from software subscriptions, AI capabilities, cloud - based CAM, creative tools, as well as content communities and process asset transactions formed around models, tool paths, and processing parameters.

In terms of the team background, the members of Infimaker come from top - tier intelligent hardware companies such as DJI, Xiaomi, Dreame, and Yunjing. Founder Xie Bowen was responsible for the perception algorithm of the sweeping robot at DJI. The company's core team has long been deeply involved in robotics, motion control, machine vision, intelligent algorithms, and consumer - grade hardware systems, and has the complete ability from underlying technology, self - developed numerical control systems to product implementation. Team members have led the R & D and mass production of multiple global intelligent hardware products and have accumulated rich experience in high - end equipment, consumer electronics, automation systems, and global brand operation.

CEO Interview:

Yingke: Why did you choose to enter the CNC track in 2024?

Xie Bowen: First, we identified the underlying demand. I started making robots since the second grade of primary school and have been a Maker for decades. In the entire personal creation chain, "manufacturing" has always been the most difficult part. Why now? Because 3D printing has completed the first wave of market education, and more and more people are starting to try to make things by themselves. However, additive manufacturing has difficulty solving the problems of multi - material and high precision, which are exactly the areas where CNC excels. Coupled with the popularization of robot technology and the reduction of creation and usage thresholds by AIGC, we believe that now is the time for personal subtractive manufacturing to truly enter the mass market.

Yingke: What specific changes did your DJI background bring to the product - defining ability?

Xie Bowen: The biggest influence of DJI on us is that a good product should not require users to think too much. From the very beginning, we thought about how users can turn an idea into a product with as few operation steps as possible. So we self - developed control algorithms and CAM, simplified the workflow, and used visual perception algorithms to reduce operation time. We don't want users to "learn the machine" but hope that creativity can be realized more simply.

Yingke: How do you view the market space and competition of desktop CNC?

Xie Bowen: Currently, the market is still in the education stage. The real challenge of consumer - grade CNC is not "making the machine", but how to enable ordinary users to process successfully and stably. We hope that more excellent teams will join this track. Our core advantages mainly lie in several aspects: First, algorithm accumulation. We cut several tons of materials to master the five - axis cutting parameters. Second, the perception algorithm and control ability of the integration of software and hardware, including 3D reconstruction, broken - tool detection, and error compensation. Finally, the ecosystem. The community we want to build not only has model data assets but also includes processing parameters that match the machines, materials, and tools. Hardware is just the starting point, and the long - term tuning of the combination of software and hardware and the precipitation of the ecosystem are walls that latecomers are difficult to cross.