Fujitsu announced the development of a Neural Network Potential (NNP) learning method for molecular dynamics simulations on December 1, 2025. According to the press release, the technology allows for the simulation of solid-state battery interfaces involving over 100,000 atoms over 10 nanoseconds with high accuracy. This is a significant leap compared to traditional methods, accelerating R&D in new materials and the creation of more efficient and safer batteries.
Fujitsu Develops Neural Network Method for Battery Simulation (>100k Atoms)
