Dunlop, a brand under Sumitomo Rubber Industries, Ltd., and Fujitsu Limited have announced the development of an AI surrogate model aimed at predicting tire performance with high accuracy and reduced analysis time. This collaboration marks a significant step in tire design digital transformation, reducing analysis time from 45 minutes to approximately 5 minutes, while maintaining accuracy.
The AI surrogate model, developed as part of Dunlop's long-term management strategy, was applied to the structural analysis of tire deformation in contact with road surfaces. Utilizing Fujitsu's AI technology and Dunlop's tire design expertise, the model is based on a Graph Neural Network algorithm, achieving an average accuracy of 87.7% compared to traditional Finite Element Method (FEM) analysis.
This new technology runs on Fujitsu's FUJITSU-MONAKA, an Arm-based CPU designed for high performance and energy efficiency. The companies plan to optimize inference speed, accuracy, and power efficiency by testing the model on a prototype of FUJITSU-MONAKA by December 2026.
The initiative aims to enable faster decision-making in tire design, reducing the number of processes required to determine tire structure and material specifications. By April 2027, Dunlop plans to implement this technology in practical operations, accelerating data-driven development and improving tire safety and environmental performance.
Fujitsu intends to apply this AI technology across other manufacturing industries, promoting carbon neutrality through enhanced power efficiency. The company will offer the AI inference platform, combining FUJITSU-MONAKA and the Graph Neural Network, via its AI platform Fujitsu Kozuchi.
The collaboration aligns with Dunlop's corporate strategy, R.I.S.E. 2035, and its vision of providing new experiential value from rubber. Fujitsu, committed to sustainability, views this development as a means to optimize development processes and enhance industry-wide efficiency.