- The jointly developed AI model has passed comprehensive acceptance testing and is now operating within Fangda Carbon’s production workflow.
- By forecasting performance and screening formulations before physical trials, the system meets cost optimization targets while enabling rapid evaluation of substitute raw materials.
- This initial deployment establishes a scalable foundation for XtalPi’s industrial AI business in new materials, converting complex manufacturing data into reusable, predictive models for advanced carbon materials.
XtalPi Holdings Limited (2228.HK) announced that a raw material selection artificial intelligence (AI) model jointly developed with Fangda Carbon New Material Co., Ltd. (600516.SH), a global leader in graphite electrodes, has successfully passed project acceptance and entered operational use. The model reliably evaluates and ranks formulation candidates prior to physical trials, meeting established targets for predictive accuracy, cost optimization, and operational efficiency.
Graphite electrodes are critical components that conduct the massive electrical currents required for electric arc furnace steelmaking. Key properties, including electrical conductivity and mechanical strength, depend on complex interactions among raw material inputs, blend ratios, and manufacturing conditions.
Because materials of the same type can vary by supplier and batch, formulations must be reassessed when inputs change. Furthermore, price volatility and supply disruptions make rapid, accurate material substitution a commercial necessity.
While expert judgment remains essential, exhaustively testing every variable combination in physical trials is prohibitively slow and expensive. To navigate this physical search space, the partners combined Fangda Carbon’s six decades of proprietary manufacturing data with XtalPi’s data engineering and algorithmic capabilities. XtalPi structured decades of dispersed production records, engineered predictive features, and deployed a hybrid system that integrates performance forecasting, optimization algorithms, and embedded expert rules.
For any fixed set of raw materials, the model now rapidly identifies blend proportions that drive down costs while meeting quality requirements. When supply availability or pricing shifts, the system evaluates alternative inputs and recommends formulation adjustments, significantly expanding Fangda Carbon’s procurement flexibility.
Validation against both independent test datasets and production trials confirmed that the model achieved the required predictive accuracy across multiple key performance indicators. Its computational speed matches the real-world pace of industrial production, aligning seamlessly with Fangda Carbon’s workflow. By screening out unsuitable options and ranking the most viable candidates, the AI narrows the focus for human experts. Final formulations are still determined by experts through necessary experimental or production validation. This “compute first, verify later” approach improves decision consistency by combining industrial expertise with data-driven prediction.
The newly accepted model serves as the first core module delivered under the partners’ graphite electrode formulation optimization project, stemming from a strategic agreement signed in 2025. By connecting XtalPi’s AI, robotic experimentation, and quantum chemistry capabilities with Fangda Carbon’s industrial leadership, the partners are converting dispersed, historical expertise into traceable digital resources. Moving forward, the two companies plan to expand the system into comprehensive formulation design and process optimization.
For XtalPi, this deployment extends its industrial delivery capabilities, establishing reusable data standards, model architectures, and delivery frameworks for complex industrial applications. Adapting this infrastructure to additional advanced carbon materials, including graphene and carbon nanotubes, lays a robust foundation for expanding XtalPi’s industrial AI business and accelerating the commercial-scale production of next-generation materials.
