Battery materials innovation is becoming a defining factor in the performance, durability, and manufacturability of advanced electric vehicles (EVs). Meeting increasingly demanding energy density, reliability, and cost targets requires moving beyond conventional trial-and-error development toward more predictive, computation-driven approaches.
Mitsumoto Kawai, Chief Engineer, Device Process Innovative Research Excellence, Honda R&D Co., Ltd.
Advances in artificial intelligence (AI), computational screening, and crystal structure prediction are reshaping how candidate materials are identified, evaluated, and translated into scalable production pathways. These tools expand the searchable materials space while improving confidence before physical validation begins.
To explore how these methods are influencing battery and power device development, Mitsumoto Kawai, Chief Engineer, Device Process Innovative Research Excellence, Honda R&D Co., Ltd., shares his perspective on AI-enabled materials discovery, stability prediction, and the practical challenges of bringing advanced materials into automotive programs.
Here is what he had to say…
How is AI changing the way new materials are identified and evaluated for EV applications?
Traditionally, materials discovery relied heavily on trial-and-error processes driven by researcher intuition and experience. However, with the introduction of AI, particularly machine learning, it’s now possible to narrow down promising materials from tens of thousands of compositional candidates in a relatively short period of time, without relying on intuition and experience.
What limitations of traditional materials discovery methods become most problematic as EV systems grow more complex?
The performance required of EV batteries is becoming increasingly demanding. To meet these requirements, material compositions may become more multi-dimensional (e.g., ternary or quaternary material systems with three or more constituent elements).
As dimensionality increases, the combinations of elements and crystal structures grow dramatically. Considering development time and cost, it is becoming increasingly difficult to solve these challenges through experimentation alone.
Advanced AI modeling is helping engineers evaluate and optimize EV battery materials before physical validation begins.
Why are multi-component and metastable materials increasingly important in EV batteries and power electronics?
Their importance is increasing because multi-component systems and metastable phases have the potential to achieve properties that are difficult to realize with materials composed of relatively few elements or with stable phases.
How can computational screening and crystal structure prediction reduce experimental iteration time in EV materials research?
Computational materials screening makes it possible to eliminate materials with low theoretical feasibility before conducting experiments. Crystal structure prediction methods calculate the thermodynamic stability of target materials. By using these results, it becomes possible to evaluate the feasibility of materials prior to synthesis, enabling experiments to be conducted more efficiently.
How do faster materials evaluation workflows affect development timelines for EV batteries and energy systems?
Shortening the material development period ultimately reduces the overall development timelines. Additionally, within the same development period, multiple material candidates can be evaluated in parallel, allowing for more flexible responses to specification changes.
Crystal structure prediction and computational screening allow engineers to assess material stability and feasibility prior to synthesis.
What challenges arise when scaling AI-driven materials discovery from research environments to automotive programs?
At the mass-production level, consistent reproducibility is required but achieving it is very difficult, particularly for multi-component systems and materials with metastable phases.
On the other hand, crystal structure prediction methods can calculate the stability ranges of metastable phases, phase transitions, and phase separation. Moving forward, a key challenge will be translating these calculated stability ranges into practical manufacturing conditions.
How might AI-enabled materials modeling influence long-term EV performance, reliability, and manufacturability?
AI-enabled materials modeling is expected to enable the rapid development of battery materials with high performance and reliability, as well as the establishment of highly reproducible manufacturing processes. This may facilitate the early market introduction of innovative battery technologies
Filed Under: AI Engineering Collective, Featured Contributions, Q&As