Case Study: Sumitomo Electric achieves 10x faster AI performance and 50% shorter materials development with Rescale

Sumitomo Electric Industries Redefines R&D Innovation with AI on Rescale

Sumitomo Electric, a global manufacturer headquartered in Osaka, faced a compute bottleneck as its new Materials Informatics (MI) and Process Informatics (PI) teams adopted AI/ML workflows (including GANs) that outstripped their on‑premises HPC. To run large-scale brute‑force screening (over 100,000 candidates) and support open‑source tools like LAMMPS, Quantum ESPRESSO and PyTorch, Sumitomo Electric turned to Rescale’s cloud HPC platform.

Rescale delivered an instantly available, secure cloud HPC SaaS with high‑memory GPUs, preinstalled scientific and AI software, and scalable storage, enabling centralized application and resource management. With Rescale, Sumitomo Electric accelerated AI workloads roughly 10x, cut computation times by up to 90% (some week‑long jobs now run in half a day), halved material development time (about 10 → 4–5 months), supported 20+ concurrent projects, and produced five patents in a year. Rescale’s platform also improved collaboration and ROI by letting teams rapidly scale resources as needed.

Sumitomo Electric

Tatsuya Takakuwa

Group Manager