In a breakthrough for materials science, NVIDIA has introduced ALCHEMI, its AI Lab for Chemistry and Materials Innovation. This AI-driven platform is designed to revolutionize material discovery, transforming the traditional process which often takes decades into a streamlined operation achievable in months.
ALCHEMI’s AI-accelerated workflow consists of four stages: hypothesis generation, solution space definition, property prediction, and experimental validation. The platform offers APIs and microservices to assist developers in deploying generative AI models and AI surrogate models, using machine learning interatomic potentials (MLIPs) to predict material properties cost-effectively.
NVIDIA’s Batched Geometry Relaxation NIM has demonstrated an 800x speedup in some scenarios, and SES AI, a lithium-metal battery technology company, is using ALCHEMI NIM microservice to accelerate the identification of new electrolyte materials. NVIDIA aims to map up to 10 billion molecules in the future, showcasing the potential of AI to drive significant breakthroughs in material science and create a more sustainable future.