Summary of Conference Call on Zhongkong Technology Company Overview - **Company**: Zhongkong Technology - **Industry**: Continuous Process Industries (Petrochemical, Chemical, Power, Metallurgy) Key Points 1. **Introduction of TPT Model**: - TPT stands for Time-series Pre-trained Transformer, developed over three years by a team of hundreds at Zhongkong Technology. It is the first pre-trained AI model system designed specifically for continuous process industries, currently at version 2.0 [1][1]. 2. **Competitive Edge of TPT Model**: - **High Barriers**: - Data barrier: Training the TPT model requires extensive factory production data, including time-series data from tens of thousands to millions of sensors per factory [1][1]. - Application barrier: Safety in chemical production is paramount, deterring inexperienced vendors from entering the market [1][1]. - **Market Position**: - Compared to large language models, the competition is fierce, but Zhongkong's TPT model has established a strong position [2][2]. 3. **Market Potential**: - In China, there are 280,000 suitable process manufacturing units, each with a subscription price of 1 million per year, leading to a potential market size of 280 billion per year [2][2]. - The product has expanded beyond China, with international clients like Saudi Aramco, indicating a larger global market opportunity [3][3]. 4. **Revenue Projections**: - TPT model revenue exceeded 200 million in 2025. The company aims for 800-1,000 million in 2026, 1.5-2.5 billion in 2027, and 3-5 billion in 2028 [3][3]. - The revenue growth rate and scale are comparable to other large model companies [4][4]. 5. **Core Business and Valuation**: - Zhongkong Technology is not just an AI application company; it possesses core barriers and significant growth potential as a large model company. The traditional DCS business still generates a stable profit of 1.2 billion, while the large model segment has only received a valuation of 200-300 billion [5][5].

Core Insights - The concept of "Industrial Embodied Intelligence" was introduced by Zhongkong Technology, aiming to integrate intelligence into production processes, transforming it from a "tool" to a "subject" with continuous learning and autonomous operation capabilities [1][3] Group 1: Implementation and Technology - Hubei Xingrui Chemical Co., a subsidiary of Xingfa Group, became the first application site for Industrial Embodied Intelligence, replacing multiple traditional control systems with a unified General Control System (UCS) that manages over 15,000 data points [2] - The UCS allows for automatic recognition of process parameter fluctuations, timely adjustments, and proactive alerts for abnormal trends, effectively creating a complete production closed loop [2] - The TPT large model performs multi-modal data analysis and trend forecasting, ensuring precise control of key product concentrations and optimizing resource consumption [2] Group 2: Industry Impact and Future Prospects - The successful implementation of Industrial Embodied Intelligence marks a shift in process industry operations from "experience-driven" to "AI real-time autonomous decision-making" [3] - TPT serves as a universal template that can be adapted across various industries with minimal adjustments, enabling rapid activation of six core capabilities including simulation, control, and optimization [3] - The application of Industrial Embodied Intelligence is expanding beyond the petrochemical sector into metallurgy, construction materials, and discrete industries, with increasing proactive engagement from clients [3]