zhangtongshe.com
The industrial embodied intelligence track is currently at a critical turning point where the hype is cooling down and practical implementation is taking the forefront.
The embodied intelligence sector is currently facing a deep divide.
In various exhibition halls and launch events, robots operate smoothly and intelligent models respond accurately. However, the real test begins when stepping into the front-line workshops of heavy industries such as shipbuilding, nuclear power, and offshore engineering. On-site working conditions are highly dynamic, and the red lines for production precision and safety are extremely strict. Embodied solutions that run perfectly in the lab are likely to struggle with implementation on complex, real-world production lines.
Between the hype of the trend and industrial implementation lies a chasm that the vast majority of players cannot cross.
On September 22, the "Smart Manufacturing in Baoshan · Embodied Future" 2026 Industrial Embodied Innovation Ecosystem Conference was held in Baoshan, Shanghai, bringing these issues to the table. One of the conference organizers, Robotic Plus. AI (Shanghai) Co., Ltd. (referred to as "Robotic Plus"), is a company researching the "industrial embodied brain." It brought scenario providers, technology providers, robot manufacturers, research institutions, and capital to the same table to jointly tackle this challenge.
The conference aimed to answer: How can embodied intelligence truly enter factories and get to work? And how exactly should the ecosystem for industrial embodied intelligence be built?
At the Conference: Facing the Real-World Dilemmas of Industrial Embodied Intelligence
This conference pushed the real-world challenges of the industry to the forefront.
At the scene, "China's No. 1 Shipyard" Shanghai Waigaoqiao Shipbuilding and China's nuclear power giant China National Nuclear Corporation (CNNC) took the stage to release real industrial scenario demands. Experts from Robotic Plus, Shanghai FANUC, and Tongji University subsequently took the stage to introduce practical implementations and cutting-edge research in industrial embodied intelligence. During the panel discussion, representatives from the manufacturing sector, robot enterprises, and guests from institutions such as Linear Capital took the stage together for in-depth cross-disciplinary exchanges.
Liu Jianfeng, Chief Expert of China State Shipbuilding Corporation (CSSC) and Chief Process Engineer of Shanghai Waigaoqiao Shipbuilding, stated frankly that shipbuilding is highly non-standard, requiring a well-built ecosystem, as going it alone leads nowhere. Wu Di, Deputy Chief Engineer of China Nuclear Engineering & Construction Corporation (CNECC) and General Manager of CNECC Chuangke, also pointed out that embodied intelligence cannot be separated from scenarios; only by clearing the pathway in advance can the technology truly be implemented.
Through the collision of dialogues from different perspectives, the industry has gradually reached several key consensuses.
First, general large models are highly capable, but industrial scenarios do not need models that "can chat"; they need models that "can work."
Second, the hardware of robot bodies is becoming increasingly mature, and costs are dropping rapidly, but there is a lack of a "robot brain" that can adapt to complex, non-standard working conditions.
Third, high-quality closed-loop industrial physical data is extremely difficult to obtain.
These consensuses point to the same judgment: the competition in industrial embodied intelligence is shifting from "who has the flashier technology" to "who can get things done in real-world scenarios."
And to truly get things done, relying on a single enterprise is far from enough.
Based precisely on this judgment, Robotic Plus took the lead in initiating the Industrial Embodied Industry Ecosystem Alliance, which was officially established at this conference. The first batch of members includes Shanghai FANUC, the State Key Laboratory of Autonomous Intelligent Unmanned Systems at Tongji University, the Joint Laboratory of Intelligent Design and Autonomous Construction at Shanghai Jiao Tong University, Haier Capital, Mech-Mind, YOUIBOT, DH-Robotics, Wuxi Xinshanlian, and Current Robotics. This alliance integrates scenario providers, robot manufacturers, technology companies, universities, and capital into a normalized collaborative mechanism.
Tongji PhD Entrepreneurship: Solving the Challenges of Industrial Embodied Intelligence
Looking at Robotic Plus's development history, the company has long addressed the challenges of non-standard customized production, bridging the data closed-loop from design to manufacturing. This is precisely the core proposition that the implementation of industrial embodied intelligence must face.
In 2016, Meng Hao, a PhD from Tongji University, founded Robotic Plus in Shanghai with partners from ETH Zurich, Zhejiang University, and the Chinese Academy of Sciences.
Over the past decade, Robotic Plus has personally experienced almost every challenge discussed at this conference.
Facing the first hurdle of industrial embodied intelligence, which is reliable processes and controllable safety, Robotic Plus plunged into the most challenging industrial sites such as cutting and welding. These processes have extremely high requirements for precision and safety; a single misstep could halt the entire production line. Currently, the company's three products—bevel cutting, welding, and 3D cutting—each have annual sales exceeding 50 million CNY. Among them, the bevel cutting robot has secured the No. 1 spot in its niche market, with annual sales of a single product exceeding 70 million CNY.
Meng Hao's positioning of Robotic Plus precisely hits the industry's second pain point.
He realized that while robot body hardware is becoming increasingly mature, the industry lacks a "robot brain" that can adapt to complex, non-standard working conditions. Therefore, instead of manufacturing robot bodies, he chose a more difficult path: developing the RoBIM industrial software platform in-house. Today, RoBIM has evolved into a cloud operating system integrating CAD, CAM, and CAPP, compatible with mainstream robotic arm brands both domestically and internationally. It integrates various industrial sensors and tool heads, quickly assembling flexible solutions like industrial Lego.
The third hurdle is that high-quality closed-loop industrial physical data is extremely difficult to obtain. Many enterprises lack real factory scenarios to iterate algorithms, whereas Robotic Plus has naturally accumulated a massive amount of real-machine data over years of delivery. From customer drawings to robot motion trajectories, camera perception, and finished product precision verification, the data of the entire machining process is deposited on the RoBIM platform. "We now have a closed-loop data set of over 100,000 real-world instances of robots cutting, assembling, and welding components in factories," said Meng Hao.
These capabilities, forged in real production lines, have enabled Robotic Plus to run through a closed loop from scenarios to data, and from data to products.
But what Meng Hao wants to do goes far beyond just making his own company better at getting the job done.
Shanghai's Embodied Intelligence Industry Accelerates Its Momentum
Today, Robotic Plus is gradually growing from a product champion into the ecosystem "chain master" of the industrial embodied intelligence track.
Industrial embodied intelligence is not an undertaking that a single enterprise can complete alone. This is a judgment that Meng Hao has repeatedly emphasized on multiple occasions.
Take this conference as an example. Shanghai FANUC, Tongji University, Linear Capital, Mech-Mind, YOUIBOT, DH-Robotics, Current Robotics, and Wuxi Xinshanlian are all leading institutions in their respective fields, and they share a common identity—partners of Robotic Plus.
Shanghai FANUC co-hosted the event and provided high-reliability robots and welding process packages, jointly developing multiple industry cases with Robotic Plus in coal mining machinery, port machinery, and shipbuilding; Tongji University and Shanghai Jiao Tong University conduct joint R&D with Robotic Plus; Linear Capital is an early investor in the company; and enterprises such as Mech-Mind, YOUIBOT, and DH-Robotics have also formed synergies with Robotic Plus in fields like 3D vision, mobile robots, and dexterous hands.
The willingness of these partners to put hardware, research, capital, and niche technologies on the same table to interface with Robotic Plus's system foundation itself demonstrates that the industrial embodied ecosystem built by Robotic Plus is rapidly taking shape.
Meanwhile, Robotic Plus's position as the ecosystem "chain master" is being repeatedly validated by the upstream and downstream of the industrial chain.
On the client side, Robotic Plus's intelligent robot solutions have entered the production lines of leading enterprises such as Jiangnan Shipyard, Waigaoqiao Shipbuilding, ZPMC, CNNC, CCTEG, Harbin Electric, and CSCEC. Nearly 1,000 sets have been sold cumulatively, serving nearly 300 leading customers globally. The two scenario providers that released demands at this event, Waigaoqiao Shipbuilding and CNNC, are also partners of Robotic Plus.
On the capital side, Robotic Plus has cumulatively secured nearly ten rounds of financing. In March this year, the company completed a Series D funding round of hundreds of millions of CNY, co-led by Bova Capital and CICC Capital. What capital values is not just commercial returns, but also Robotic Plus's technological barriers and industrial position in the industrial embodied intelligence track.
In 2025, the company was awarded the title of National Key "Little Giant" Enterprise. This is not just a recognition of the enterprise's scale, but also an official confirmation of Robotic Plus's position in the industrial embodied intelligence industry chain.
Looking to the Future: Domestic Industrial Embodied Intelligence Ushers in New Opportunities
The industrial embodied intelligence track is currently at a critical turning point where the hype is cooling down and practical implementation is taking the forefront.
The national "15th Five-Year Plan" has elevated embodied intelligence and industrial software to a strategic height. Shanghai has proposed striving to push 100,000 humanoid robots into factories by the end of the "15th Five-Year Plan," with the penetration rate of intelligent agent applications in industrial enterprises above designated size exceeding 80%. As Shanghai's leading area for intelligent manufacturing, Baoshan District has gathered 332 robot-related enterprises, achieving an industrial output value of 32.3 billion CNY for industrial enterprises above designated size in the first half of the year, covering nine niche tracks including industrial robots, embodied intelligent robots, and industrial mother machines.
In this era, whoever can truly solve non-standard working conditions, whoever can turn data into productivity, and whoever can make the ecosystem operate synergistically will take the initiative in the next stage.
Today, Robotic Plus's growth curve perfectly aligns with this mainline of policy and industrial resonance.
Based on the real process data accumulated by RoBIM, the company is extending the adaptive capabilities of industrial robots to collaborative robots, and even providing an implementation foundation for future humanoid robots in industrial scenarios. At the same time, Meng Hao is also planning overseas expansion to push Chinese products to the world. From energy to electric power, aerospace, and offshore engineering, Robotic Plus is stepping onto a larger stage.
As more and more robots enter factories and head to global infrastructure sites, the ultimate vision Robotic Plus wants to achieve has become clear.
To ensure that the efficient operation of every industrial robot is driven by a domestic industrial intelligence foundation. To make China's industrial embodied technology a core force in global manufacturing upgrades.
Text | Na Yun Layout | Guo Beibei