In 2025, SUPCON Technology's new product, TPT 2, was officially launched globally.

On August 28, the 2025 SUPCON Global New Product Launch and Industrial AI Innovation Development Conference, themed "Harnessing Temporal Energy, Guiding the Future with Wisdom," was successfully held in Hangzhou. SUPCON officially launched the Time-series Pre-trained Transformer 2 (TPT 2) to the global market.

The conference attracted nearly 700 representatives from leading industrial enterprises, multinational organizations, industry associations, research institutes, investment institutions, and well - known media from dozens of countries and regions around the world. This launch event focused on the in - depth integration of process industry and artificial intelligence, and conducted in - depth discussions on the technological changes and development momentum brought by industrial AI. The event was broadcast live globally on multiple platforms such as Xinhua Net, Sina Finance, 36kr, and Gongkong.com. The online live - viewership reached nearly 7 million, the reading volume of the Weibo topic exceeded 140 million, the discussion volume reached over 85,000, and the interaction volume exceeded 97,000. It ranked 40th on the Weibo hot - search list and 5th on the Weibo technology list, attracting wide attention.

Global Premiere of TPT 2

Zhang Huize, the vice - president of SUPCON's international operation company, emphasized during the new product launch: "TPT 2 is a revolutionary industrial AI tool specifically designed for the core needs of the process industry."

This model builds an AI algorithm based on industrial time - series data and is a credible model that deeply conforms to the first - principles of the process industry. With strong generalization ability, it can be widely adapted to various production processes and achieve closed - loop applications. It can accurately tap the value of industrial data and play a core role in key scenarios such as ensuring production safety, improving product quality, reducing energy and material consumption, optimizing equipment operation and maintenance efficiency, and promoting the autonomous operation of factories. It provides integrated support for enterprises from solutions to industrial agents.

Specifically, the core value of TPT 2 is reflected in three dimensions:

Intensive Empowerment, Penetrating the Data Value Layer

Relying on the SCOPES capability matrix (Simulation, Control, Optimization, Prediction, Evaluation, Statistics), TPT 2 subverts the traditional decentralized development model of "one model for one scenario". Through language interaction, it can efficiently generate agents applicable to various process units, replacing N industrial software in simulation, control, optimization, prediction, evaluation, statistics and other links, achieving intensive intelligent support and efficiently releasing the deep - seated value of data.

Expert - level Interaction, Real - time Anchoring of Decision - making Points

By deeply integrating the laws of industrial production data, TPT 2 can achieve root - cause positioning and accurate prediction, opening up a new path for diagnosing industrial problems. With conversational interaction, users can easily analyze and handle production problems, enjoying 7x24 - hour online expert - level decision - making support, which directly enhances the core competitiveness of enterprises.

Full - link Closed - loop, Breaking through the Bottleneck of Unmanned Operation

With the full - link closed - loop ability of "perception - recognition - decision - execution", TPT 2 can overcome the limitations of equipment and scenarios, cope with complex industrial environments, and provide a solid technological foundation for the transformation of enterprises towards less - manned, unmanned, and highly intelligent operations.

As an industrial agent platform centered on MoE (Mixture of Experts), TPT 2 deeply integrates multiple technical systems, promoting the evolution of industrial applications from traditional tool - type assistants to agents with the ability of active anomaly recognition, intelligent risk assessment, and autonomous decision - making and execution. It reshapes the production paradigm with scenario - based intelligent solutions, achieving significant innovation in the technical system and application mode of industrial software.

Subsequently, Zhou Yexu, an algorithm development engineer at SUPCON, Lin Xiang, a product engineer, and Zeng Xinxin, a research and development engineer, respectively shared their insights on TPT 2 in terms of time - series data algorithms, human - machine interaction and collaboration, and reliability verification.

TPT 2 Helps Enterprises Compete in the Market with Lower Costs and Higher Quality

The launch of TPT 2 is a historical turning point in the development of industrial AI. It breaks not only the functional boundaries of traditional industrial software but also the century - old industrial inertia of "experience - driven production".

Chu Jian, the founder of SUPCON, pointed out in his speech that based on the verified usability and practicality of TPT 1 in multiple cases, the newly launched TPT 2 with a newer architecture and a stronger model further promotes the intelligent application of the process industry. He reviewed the evolution of industrial control systems from DCS to UCS over the past 30 years, emphasizing that the leap in network technology has brought about disruptive progress to the industry, and pointed out that generative AI and general AI are accelerating industrial transformation.

Chu Jian said that TPT 2 is not simply an application of general large - scale models but deeply integrates industrial scenarios, aiming to solve core problems such as safety, quality, cost, efficiency, and low - carbon emissions that enterprises generally care about. He hopes that by improving production safety, product quality, and energy efficiency, it will help enterprises enhance their competitiveness in the current economic environment. In addition, SUPCON will continue to iterate TPT, strengthen cooperation with users, and promote Chinese industrial AI to have an important influence globally, making it a core tool to support the intelligent transformation of the industry.

Industrial AI Data Alliance: Building a Data - value Ecosystem Together

During the conference, the founding ceremony of the "Industrial AI Data Alliance" was grandly held. SUPCON, with an open attitude, initiated an initiative and joined hands with more than 130 industrial ecological partners, including leading enterprises, design institutes, general contractors, and service providers in the industry, to jointly build the "cornerstone of trust" for the development of industrial AI. This cross - industry and open - ended organization will be committed to promoting the scenario - based and large - scale application of industrial data and service sharing, empowering the intelligent transformation and sustainable development of the process industry.

"Linking Industry with Data, Gathering Wisdom in the Nebula", the alliance will focus on three major directions: data sharing, value co - creation, and ecological co - construction:

• By building a trusted industry data space and an industrial data standardization governance system, it will break the dilemma of "fragmented data, unusable data, and insufficient collaboration" in the industrial field, accelerating the transformation of data into value.
• On the premise of ensuring data ownership and privacy security, it will realize the circulation of device data across scenarios and working conditions, providing high - quality "nutrients" for the pre - training of industrial large - scale models and empowering enterprises' intelligent upgrading.
• It will continuously promote the R & D of industrial AI and scenario incubation, deepen the collaboration and upgrading of the industrial chain, build a resource hub and a value network, and achieve a positive cycle of "resource gathering - value co - creation - benefit sharing".

Yang Jianting, the deputy general manager of Lanzhou Petrochemical Company of PetroChina, Liu Qiang, the member of the Party Committee and deputy general manager of Shandong Energy Group Co., Ltd., Sun Shaowen, the general manager of Wanhua Chemical (Ningbo) Co., Ltd., Wang Cong, the deputy general manager and chief engineer of China Tianchen Engineering Corporation, Luan Bo, the executive director and chief rotating president of Shandong Jingbo Holdings Group Co., Ltd., Liu Wanzhou, the chairman of Inner Mongolia Yitai Chemical Co., Ltd., Xu Lie, the chairman of Huataiyongchuang (Beijing) Technology Co., Ltd., Yang Fang, the general manager of Shanghai Huayi Information Technology Co., Ltd., together with Cui Shan, the chairman and president of SUPCON Technology Co., Ltd., and Wu Yucheng, the vice - president, jointly launched the founding ceremony of the alliance.

Representatives from all walks of life, including Pang Guanglian, the member of the Party Committee, chairman and secretary - general of the Foreign Investment Committee of the China Petroleum and Chemical Industry Federation, Guan Qingyou, a famous economist, the dean of Rushi Financial Research Institute, and the founding partner of Rushi Capital, Huang Pei, a member of the National Expert Committee on Intelligent Manufacturing, Liu Yuhui, a member of the China Chief Economist Forum and the deputy director of the Shanghai Chief Economist Financial Development Center, Gu Jiandang, the president of Phoenix Contact China, Zhao Yuning, the executive president of Zhejiang Dahua Technology Co., Ltd., and Chu Jian, the founder of SUPCON, the initiator of the alliance, witnessed the event on - site.

Guest Sharing: Valuable Insights

In the keynote report session, Li Wei, the general manager and chief engineer of Lanzhou Petrochemical, Yao Bohan, a student representative from China University of Petroleum (East China), and Yin Ye, the CEO of BGI Group, respectively shared their in - depth and multi - angle insights on the practical results of TPT 2 in industrial scenarios and the development prospects of AI technologies such as large - scale models in the industry.

Li Wei, General Manager and Chief Engineer of Lanzhou Petrochemical of PetroChina

SUPCON's time - series large - scale model TPT has been successfully applied in the Yulin Chemical Project of Lanzhou Petrochemical of PetroChina, effectively addressing industry problems such as weak perception ability of ethylene plants, difficult benefit balance, and slow operation adjustment. Three core agents built on industrial AI algorithms play a key role: The anomaly early - warning agent monitors production parameters and equipment status in real - time, with a prediction accuracy of 99.79%. It can accurately locate anomalies and generate disposal plans to ensure safe operation; The operation path planning agent optimizes the COT temperature operation path during the furnace - charging and heating - up stage, shortening the heating - up time by 4 - 5 hours. On the premise of meeting quality standards, the overall annual benefit is expected to be no less than 6.5 million yuan; The operation optimization agent improves the selectivity and yield of ethylene during the stable operation stage, increasing the ethylene yield per furnace by 0.373%, with an annual net income of 3.155 million yuan per furnace. The overall annual benefit is expected to be no less than 15 million yuan.

Yao Bohan from China University of Petroleum (East China)

Yao Bohan, a student representative from China University of Petroleum (East China), shared his experiences and expectations under the theme of "New Forces in Industrial AI: Practice and Expectations from the Competition Field to the Industry". As a key member of the winning team in the recent "SUPCON Cup" university competition, he mentioned in his speech that when facing complex problems in industrial scenarios during the competition, traditional software systems often showed many limitations. In contrast, TPT 2 is truly "easy to use and useful", which can help him achieve twice the result with half the effort in project research and allow him to experience the real - sense of actual factory production operations during his school days. In the future, when he takes up relevant positions, it will enable him to handle production operations and optimize complex situations more efficiently.

Yin Ye, CEO of BGI Group

Mr. Yin Ye, the CEO of BGI Group and a biology doctor, put forward in his report titled "Timing with Rhythm, Manufacturing with Wisdom - TPT 2: Awakening the Industrial Pulse with Life Wisdom": True intelligence should be based on the understanding and respect for the basic rhythm of life - 'timing'.

After billions of years of evolution, natural systems have formed a highly adaptable and dynamically balanced operating mechanism, which inspires industrial intelligence to go beyond the mechanical maximization of efficiency and move towards a wise integration that resonates with the rhythm of life. With advanced algorithms such as timing perception, artificial intelligence can perceive the production rhythm, predict fluctuations, and respond dynamically like a living organism, achieving a deeper level of collaboration between equipment and humans. TPT 2 represents this cognitive reconstruction: regarding the industrial system as an organism that can breathe, learn, and grow, making faults predictable and bottlenecks malleable, and ultimately forming a