Long before artificial intelligence became a mainstream topic, global brake friction manufacturer TMD Friction was pioneering its use to develop new friction materials more efficiently. While many view AI as a recent phenomenon, the company has been integrating this technology into its Research & Development (R&D) processes for nearly a decade, viewing it as a powerful evolution of data management that complements, rather than replaces, human intelligence.

Key Highlights

TMD Friction has used sophisticated, in-house AI tools to develop new friction formulas since 2015.

AI-supported compounding and “virtual testing” make the R&D process more efficient, significantly reducing the time and resources spent on dynamometer testing.

The company’s advanced AI can predict how changing the concentration of a single raw material will impact the final product’s performance metrics.

While AI provides powerful data analysis, human oversight remains essential for final product decisions, accounting for market, economic, and legal factors.

A History of Technological Advancement

TMD Friction’s journey with AI began long before the technology became widely available to consumers. As early as 2004, the company was using early predecessors to AI to predict the properties of certain material mixtures. A significant leap occurred in 2015 with the introduction of AI-supported compounding, which allowed the system to predict the influence of specific raw materials on a formula’s performance. For the first time, the technology could synthesize novel mixtures that human compounders might not have conceived of on their own.

This innovation led to the development of “virtual testing,” an AI-driven process that can predict the behavior of entirely new and untested formulas. Given that a single physical test run can cost thousands of euros and take up to a week and a half, virtual tests are instrumental in saving critical resources. By 2020, TMD Friction further refined its technology, training an AI model to determine precisely how adjusting the quantity of an ingredient would affect performance, making it the first automotive supplier to integrate this level of AI into its R&D.

The BRAKE ReportJune 27, 2025Automotive