When “full-stack self-research” becomes an industry buzzword and “Integrated Drive & Control” (IDC) shifts from an elective to a mandatory challenge—do we truly understand the weight behind these terms?

In this episode of ATOMROBOT Trialogue, we dive deep into the neural center of industrial robots. We invited two technical commanders who personally “answered the exam”—Shao Lingxuan and Chi Pengqing, heads of ATOMROBOT’s IDC department—for a candid conversation about technical depth, difficult choices, and reconstruction.

This is an inquiry into the fundamental “Why”:

While others are “decorating,” why did we insist on “building the foundation”?

When integrated solutions were readily available, why did we choose the most exhausting path?
 

The Choice: “Decorating” or “Building the Foundation”?

“The ‘self-research’ many competitors talk about is, on the software side, just a secondary encapsulation of open-source frameworks—the bottom layer still belongs to someone else. On the hardware side, it’s just integration: sourcing plus assembly.”

Chi Pengqing’s words cut like a scalpel through the ambiguity of the term “self-research” in the industry.

Atomrobot’s “full-stack self-research” means everything—from bottom-layer algorithms and hardware architecture to top-layer applications—is written and built in-house.

Why take the harder road?

“Because the path of integration is ‘using whatever is available,’ while the path of self-research is ‘creating whatever is needed,’” Chi explains.

The Transition: When “No One to Blame” Becomes the Best Teacher

Full-stack self-research has one unique “benefit”: there is no one else to blame—it’s all on you.

Shao Lingxuan recalls early debugging sessions: individual modules worked fine, but as soon as they were connected, the system would “glitch,” showing unexplainable distortions in dynamic response. Mechanical, electrical, and software teams pointed fingers at each other, eventually tracing it back to embedded performance issues.

“In the past, resources were abundant, so we never considered internal module allocation. On an embedded platform, the software must be designed to manage resources based on the specific hardware constraints.”

Chi Pengqing also faced internal pushback: “Why take such a huge risk when there are ready-made solutions on the market?”

The answer that silenced the skeptics was a very real market pain point: Cost.

“The industry is too competitive. Only with controllable costs can we survive in the long run.”

When projects delayed or performance stalled, did management provide resources or “slam the table” to demand a switch to third-party solutions?

The answer: They did both.

“Giving resources was true support; slamming the table was to push the schedule, not to change the plan,” says Chi.

Shao adds: “The table-slamming had value too. Once, after the CTO slammed the table, he coordinated resources the very next day to help us solve the problem.”

Reconstruction: From “I Can Only Do This” to “I Can Solve This”

Once they mastered Integrated Drive & Control, the way they defined products changed.

Chi Pengqing explains the shift: “Before, making a product was about ‘this is the best I can do with what I have.’ Now, if there’s a problem, I have the power to solve it.”

Take the closed-loop control between the controller and the servo. Previously, when buying third-party drivers, you could only interact with fixed parameters. “It was very hard to get others to change for you. Now? You can ask for whatever you need because we are all one team.”

This shift even changed Atomrobot’s relationship with its suppliers.

“We used to ‘choose’ suppliers; now we ‘define’ them. We tell manufacturers exactly what characteristics we need, and they build accordingly. This ‘reverse customization’ gives us the initiative in the supply chain.”

What does this mean for customers?

Smaller footprint, lower cost, higher stability, unified interfaces, and support for customization. “Customers aren’t just buying a product; they are choosing a reliable partner that can adapt and grow with them long-term.”

Vision: When “Selling Shovels” Becomes a Reality

Now that the IDC technology has been conquered, it’s not just for Atomrobot’s own machines—it’s being sold as a standalone product.

“Some customers build their own robot bodies. They don’t want to build the bottom layer from scratch but still want high performance. We become their ‘technical partner,’” says Shao Lingxuan. This move is both a natural progression and a strategic layout.

When asked what the next “technical heart” will be now that IDC is becoming an industry standard:

Chi Pengqing points to the revolution sparked by Large Models (LLMs): “Imagine what AI + Robotics can do—like using AI to generate programs based on application scenarios to boost debugging efficiency.”

Shao Lingxuan values the experience of building from 0 to 1: “No matter how technology changes, as long as we have this methodology, we can re-understand and re-define anything.”

Final Thoughts: Atomrobot’s Most Precious Asset

“The most precious thing left to Atomrobot, I believe, is the courage to ‘open the black box.’ In the future, no matter how hard the problem, we won’t ask ‘can we buy it?’ first; we will ask ‘can we build it?’ This confidence is worth more than any patent,” says Chi Pengqing with conviction.

“It’s a methodology from ‘problem to answer.’ As long as this remains, we can define the future,” Shao Lingxuan concludes optimistically.

In an era where the supply chain offers everything, a company’s competitiveness has shifted from “what can you choose” to “what can you re-imagine.”

The deepest moat isn’t how firmly you stand on an existing map, but whether you have the ability to draw a new map yourself.