Click the blue text above to follow me👆
1. Who is Deng Taihua? Huawei’s “Technical Evangelist” for 20 Years If Huawei is likened to a giant ship, Deng Taihua is the engineer responsible for its core power system.
He entered the University of Electronic Science and Technology in 1995 to study communications and directly joined Huawei after graduation, starting from grassroots R&D and continuing for over 20 years. Early on: The “Invisible Force” in Communication Base Stations In the early 2000s, Huawei was in the midst of tackling 3G/4G technologies. Deng Taihua was assigned to the wireless product line, responsible for the R&D of core modules for base stations.
At that time, the outside world had only heard of “Huawei 5G,” but few knew that many algorithms ensuring stable signal transmission in base stations came from his leading team.
As one of his colleagues put it: “When users watch videos without buffering, he is happier than anyone else.” After 2010, he was promoted to Vice President of the Northeast Europe region, taking Huawei’s equipment “overseas.”
This experience transformed him from a technical expert into a “market-savvy product person”:
He had to persuade European operators to accept Huawei’s solutions while coordinating domestic R&D teams to respond quickly to demands. Mid-term: The “Ecosystem Architect” of the AI Era In 2019, Huawei launched the Ascend AI chip, and Deng Taihua was recalled to headquarters to serve as the President of the Computing Product Line.
During this period, he accomplished two key tasks: 1. Building a Developer Alliance:
The “Ascend Developer Program” he led attracted over 2 million developers within two years, effectively opening Huawei’s AI toolkit for free to the entire industry.
2. Making Technology “Grounded”:
He required the team to reduce the adaptation costs of the Ascend chip.
For example, collaborating with universities to teach students how to use Ascend for image recognition;
And partnering with companies to embed AI algorithms into industrial quality inspection equipment. At this point, Deng Taihua had transformed from a purely technical person into a “connector”:
He understood the underlying technology and could turn it into products usable by the industry. 2. The “Departure” and Reunion of ZhiHui Jun At the end of 2022, Peng Zhihui announced his departure from Huawei on Weibo, prompting mixed reactions in the comments:
“Can a popular UP master really do robotics?”
But Deng Taihua recognized the weight of this young man. From “Genius Youth” to “Wild Iron Man” Peng Zhihui’s resume is quite special:
He won over 40 science and technology competition awards during university, and after graduating with a master’s degree, he joined OPPO to work on AI algorithms. In 2020, he was recruited by Huawei under the “Genius Youth” program to join the Ascend department.
During his time at Huawei, he accomplished two notable feats: creating a mini Linux computer using components the size of a coin;
And modifying a bicycle into an autonomous driving prototype capable of avoiding obstacles to deliver packages. These works, posted on Bilibili, made him a “Wild Iron Man” with millions of fans.
However, what Deng Taihua valued most was his performance in the Ascend department:
He could optimize chip computing power using mathematical models and write complex algorithms into easily understandable code examples. Entrepreneurial Invitation: From “Technical Ideal” to “Realization” In 2022, Peng Zhihui began to consider leaving.
He later revealed in an interview: “Robots need rapid iteration, but large companies prefer long-term fundamental research.”
At this time, Deng Taihua approached him: “Why don’t we work together?”
Their roles quickly became clear: Peng Zhihui would focus on technical breakthroughs, while Deng Taihua would handle commercial realization. Someone asked: “Why Deng Taihua?”
The answer lies in the past of the Ascend department:
When Peng Zhihui’s project needed computing resources, Deng Taihua approved it;
When he wanted to create an autonomous driving bicycle, Deng Taihua coordinated the experimental site.
In Peng Zhihui’s words: “He believes in the potential of robots even more than I do.” 3. The “Unconventional Path” of ZhiYuan Robotics After ZhiYuan was established in February 2023, expectations from the outside world labeled it as the “Chinese version of Boston Dynamics.”
However, Deng Taihua and Peng Zhihui chose an “unconventional” route. Technical Route: Not Creating “Toy Robots” Their first robot, the “Expedition A1,” does not pursue flashy backflips but focuses on industrial scenarios.
For instance, in factories, it can carry 10-kilogram parts along the production line and use its robotic arm to open valves.
Deng Taihua explained: “What enterprises need is a helper that can work, not a ‘performing acrobat.'” Financing Strategy: Not Relying on “Storytelling,” but on “Networking” In just over two years, ZhiYuan completed nine rounds of financing, backed by giants like Hillhouse, Sequoia, BYD, and Tencent.
However, the most notable operation was their joint ventures with upstream and downstream companies: establishing Xier Robotics with Wolong Electric Drive to produce servo motors;
Forming a joint venture with Gree to create Zhidin Robotics, focusing on mobile chassis;
And even collaborating with the cultural tourism company Dafeng Industrial to develop a robot theater. These actions may seem scattered but are actually weaving a network:
From components to application scenarios, ZhiYuan aims to bind all key nodes of the robotics industry chain to its “war chariot.” Capital Operations: The “Triple Account” of Acquiring Shangwei New Materials In July 2025, ZhiYuan acquired 63.62% of Shangwei New Materials for 2.1 billion yuan, marking the first acquisition case of an embodied intelligent enterprise on the Sci-Tech Innovation Board.
This transaction can be considered from three perspectives: 1. Financial Account:
Shangwei New Materials had 1 billion yuan in cash, so after the acquisition, ZhiYuan effectively “spent 2.1 billion to buy an asset package and still get cash back”;
2. Qualification Account:
Although the backdoor listing was denied, the identity of a listed company gives ZhiYuan more leverage in procurement and bidding;
3. Industry Account:
Shangwei New Materials’ wind power materials business may seem unrelated, but its customer resources can open up space for robots in energy scenarios. The market responded to this transaction with stock price increases: Shangwei New Materials’ stock surged 60% after resuming trading.
One investor joked:
“They bought a new materials company, and the major shareholder became a robotics company. Who would dare to write this script?” 4. The “New Species” of the Robotics Industry While the outside world debates whether ZhiYuan is “technology-driven” or “capital-driven,” it has quietly changed the industry’s gameplay. From “Single Point Breakthrough” to “Ecosystem Bundling” Traditional robotics companies tend to focus solely on R&D, while ZhiYuan is keen on “building good relationships.”
They collaborate with Lens Technology on screen modules, jointly develop battery packs with BYD, and even co-build production lines with Luxshare Precision.
This “you have me, I have you” relationship significantly reduces supply chain risks, as no one wants to cut off their own suppliers. From “Hardware Thinking” to “Software + Services” ZhiYuan’s most underrated layout is their embodied foundational model “Qiyuan.”
This model is not just about making robots “understand human language” but enabling them to make autonomous decisions in complex scenarios.
For example, in logistics warehouses, robots can automatically adjust their handling posture based on the shape of the goods;
In research scenarios, they can even assist in completing experimental operations.
Deng Taihua’s explanation is very pragmatic:
“AI is not about showmanship; it’s about reducing errors made by robots.” From “Going Solo” to “Mixed Reform of State Capital” ZhiYuan’s shareholder list prominently features state-owned investors like Shanghai Guotou and Lingang Group.
This “private technology + state capital endorsement” model has allowed them to win bids repeatedly in local government tenders.
Industry insiders have observed:
“They are not selling robots; they are helping local governments build intelligent manufacturing ecosystems.” 5. Controversies and Challenges Despite ZhiYuan’s momentum, doubts have never ceased. “Technical Internet Celebrity” or “Real Skills”? Some have dug up Peng Zhihui’s Bilibili videos, questioning whether he treats robots as toys.
However, ZhiYuan’s disclosed data indicates that their shipment volume is expected to reach thousands of units by 2025, with clients including leading companies like CATL and SAIC.
A procurement manager in the manufacturing industry privately stated:
“A robot can replace three workers, and the cost can be recovered in three years, which is more practical than telling AI stories.” “Capital Game” or “Deep Industry Cultivation”? In April 2025, ZhiYuan announced five joint venture plans at once, which was criticized as “spreading too thin.”
However, a closer look at these collaborations reveals: Working with Junpu Intelligent on automotive assembly robots, and with Ananda to develop mobile chassis, each project is closely tied to the core of robotics.
Deng Taihua’s statement in an internal meeting was:
“We are not looking to take control but to have partners turn resources into common technical standards.” The pressure behind the rumors of “backdoor listing” Despite ZhiYuan’s repeated denials of backdoor listing, the capital market continues to bet on their indirect listing.
Investment bankers analyze:
“After acquiring Shangwei New Materials, they are already at the doorstep of the capital market. The next step is to decide whether to pursue a separate listing on the Sci-Tech Innovation Board or accelerate integration through acquisitions, depending on policy direction.” Non-Typical Sample of the Robotics Industry Looking back, the combination of Deng Taihua and Peng Zhihui resembles a metaphor for the tech industry:
The pragmatism of veterans and the fantasies of newcomers, the accumulation of traditional manufacturing and the impact of the AI wave, are wonderfully fused in them. ZhiYuan Robotics may not be the coolest robotics company, but it is proving one thing:
In the tide of manufacturing transformation, what matters more than “showmanship” is making robots truly enter factory workshops, logistics warehouses, and research laboratories.
This may be what Deng Taihua often says:
“We are not building future machines; we are solving today’s real problems.” As for whether this entrepreneurial experiment of “Huawei veteran + technical internet celebrity” will succeed?
The answer may lie in the next robot sold:
When it quietly replaces human labor in repetitive tasks, the true transformation of this era will have just begun.
(End of article)