63-year-old Chu Jian is seizing the opportunity presented by the rise of industrial intelligence. On January 6th, Supcon Technology, which he effectively controls, released an investor relations summary. In response to winning a bid for a pharmaceutical project, management revealed that they are “continuing to deeply integrate AI technology into key industrial scenarios.” Supcon Technology, located in Hangzhou, is a leading domestic industrial automation company with a market capitalization exceeding 40 billion yuan.
Chu Jian’s team has accelerated its shift towards the industrial AI field. At the performance briefing in October 2025, Supcon Technology disclosed a series of astonishing figures. Management plans to achieve annual revenue of 20 billion yuan for AI business and 5 billion yuan for robotics business within the next 4-5 years, and will invest another 20 billion yuan in AI.
</ … Its ultimate vision is to make Supcon reach a market value of one trillion yuan. Chu Jian, a native of Chun'an, Zhejiang, was formerly the vice president of Zhejiang University. He founded Supcon and broke the international giants' monopoly on DCS (Distributed Control Systems). At Supcon, this key figure only serves as a "strategic and technical advisor," with daily operations handled by Chairman and President Cui Shan and others. However, Chu remains active in the public eye, fully promoting Supcon's transformation into an industrial AI company. Chu Jian's new "weapon" is called TPT. The TPT large-scale time-series model for process industries, released in June 2024, is the core product of Supcon's AI business, drawing on over 30 years of experience in this field. Almost everything in industry is related to time-series data. For example, if a piece of equipment fails, tracing the vibration data from a week ago will almost certainly reveal the problem. If this type of data can be decoded, even predicting it a few seconds in advance can significantly improve productivity. Following the approach of large language models, Chu Jian's team pre-trained on massive amounts of industrial time-series data to learn and form general rules for the operation of industrial equipment, extracting commonalities to obtain a general basic model. Users only need to input a small amount of industry data for fine-tuning, and it can be used in various industry scenarios, such as factory operation evaluation and operation optimization. If process experts were to sort out the production logic and analyze the optimization space, it would take 1-2 months. According to Chu Jian and others' vision, using TPT to automatically decompose and calculate, the solution can be obtained in just tens of minutes, improving efficiency by more than 90%. Thus, the problem that industrial AI can only perform single-scenario customized development and is difficult to scale is solved at once. "The industrial track is very complex, and the barriers to entry are very high." A senior person in manufacturing informatization said that after the large model is developed, the technical and economic feasibility of the solution is greatly improved. In August 2025, TPT released version 2.0. Compared with the previous generation, which required experts to implement on-site, version 2.0 retains the demand for local deployment by enterprises, while adding platformization and self-service, and moving to online services to expand the customer base. On the day of TPT 2's release, 518 companies placed pre-orders, and 112 sets were sold. "This model has entered the stage of large-scale application and implementation." According to management, as revealed last October, demand for TPT far exceeded expectations, and revenue from this segment is projected to double or more in recent years.
TPT alone cannot support Chu Jian’s industrial AI transformation concept. According to his plan, Supcon aims to become a professional AI service provider in the industrial process field. Supcon has over 35,000 industrial customers. To transform AI from a supporting role to a leading one, deeply intervening in production processes and even completely taking over factories, management has introduced the concept of “industrial embodied intelligence.” “It can not only understand and judge, but also execute directly,” Cui Shan mentioned. The core of industrial embodied intelligence lies in embedding intelligence into the production process, forming a closed loop. It requires the integration of two technologies: Time Series Model (TPT) and Universal Control System (UCS), which respectively coordinate decision-making and execution capabilities. As the “physical hands and feet” of industrial intelligence, the UCS can perform precise control based on TPT’s large-scale model instructions. Supcon claims that the UCS overturns the traditional DCS architecture that has persisted for nearly 50 years. With the UCS’s minimalist structure of “one factory, one control cabinet,” cabinet space can be reduced by 90%, cable costs by 80%, and project construction time shortened by 50%. TPT + UCS, building the fully autonomous factory system FAP, will, according to Supcon’s management, drive the factory operation paradigm from primarily human-controlled to intelligently controlled. Management used an analogy: traditional digital upgrades are like a slow train, while FAP is like a high-speed train—intelligently scheduled and capable of rapid, high-speed operation. In late October 2025, Supcon secured the Henan Jinhai Fluorosilicon New Materials project based on its FAP, a project with a total estimated investment of 30 billion yuan. Regarding the Sinopec polyolefin R&D AI demonstration project won in the same period, Supcon stated, “This marks a breakthrough for TPT in the core process R&D环节 of the petrochemical field.” Chu Jian is full of confidence in the development of industrial AI. He hopes that Supcon’s customer base can soon reach 50,000. The new target set by management is indeed ambitious, aiming for a market capitalization of one trillion yuan. According to the minutes, Supcon will invest 20 billion yuan in the next 4-5 years to develop its AI business, focusing on computing power infrastructure and talent acquisition. Transforming from a software vendor to a service provider, from doing projects and products to building a platform, Supcon no longer wants to rely solely on bidding for individual projects. In terms of business model, it will shift to a membership subscription system, redesigning industrial software configurations and pricing. By the end of June 2025, Supcon had accumulated 938 subscription customers. This shift in pace will inevitably bring pressure. From January to September 2025, Supcon’s revenue and net profit both saw double-digit year-on-year declines. Chu Jian’s much-anticipated robotics business only generated 120 million yuan in revenue in nine months, failing to achieve significant scale. The pressure was perhaps predictable. Chu Jian recalled that in 2024, when Supcon launched TPT and UCS, competitors were surprised, “They didn’t expect us to overturn our own industrial software and DCS.” According to MIR DATABANK data, in 2024, Supcon Technology’s DCS market share reached 40.4%. However, the Chinese DCS market size is only about 12 billion yuan, with its ceiling clearly visible. Chu Jian predicts that in the future, a large number of application-oriented software will be replaced by AI, but engineering software involving deep scientific knowledge will still have value. With the new business segment’s foundation unstable and the old pillars replaced, Supcon still needs to find ways to prove its strength.
