SES AI Partners with NVIDIA to Accelerate EV Battery Innovation Using AI and Accelerated Computing
SES AI has partnered with NVIDIA to revolutionize the discovery and development of next-generation electric vehicle (EV) battery materials through artificial intelligence and accelerated computing.
The collaboration aims to overcome key challenges facing the EV industry, such as battery cost, range, and longevity, by speeding up the identification of advanced electrolytes, a critical component in battery performance and safety.
Using NVIDIA’s cutting-edge hardware and software platform, SES AI is leveraging domain-specific large language models (LLMs), GPU-powered molecular simulations, and AI-driven data analysis to compress what used to take decades of battery research into just a few months.
AI-Powered Approach to Battery Discovery
The initiative is structured around three core pillars: mapping the molecular universe, enhancing chemistry LLMs, and building high-performance batteries.
At the foundation of SES AI’s work is a vast molecular database built using AIMNet2, a neural network developed by Carnegie Mellon University’s Isayev Lab. This tool improves the speed and precision of molecular simulations, enabling researchers to predict the chemical and physical properties of millions of small molecules relevant to battery chemistry.
Also read: NVIDIA Unveils Isaac GR00T N1, First Open Humanoid Robot Model
Chemistry LLMs and High-Speed Simulations
SES AI trained a chemistry-focused LLM using NVIDIA’s DGX Cloud and NeMo, processing 35B tokens from 17M papers. With data refined via NeMo Curator, the model outperforms general-purpose versions in identifying and ranking promising battery materials.
It uses NVIDIA cuML with UMAP and HDBSCAN to analyze chemical datasets 10 times faster, helping identify promising battery materials. Of 17 candidates found, two are already in testing, showing a 20% improvement in battery cycle life.
Real-World Impact and Industry Adoption
The partnership has led to advanced lithium metal and high-silicon batteries, prompting SES AI to secure deals with two major automakers to accelerate next-gen battery commercialization. Where battery electrolyte discovery once took five years, SES AI’s AI-powered platform, built with NVIDIA’s accelerated computing stack, produces groundbreaking results every six months.
This rapid progress is poised to reshape the EV landscape, offering safer, more efficient batteries supporting the global sustainability push.
