SANTA CLARA — At the Supercomputing 2024 conference in Atlanta, NVIDIA introduced the ALCHEMI NIM microservice, a groundbreaking tool designed to optimize AI inference for chemical simulations. This innovation is set to transform the discovery of novel materials, advancing the renewable energy transition while addressing critical global challenges such as energy storage and environmental sustainability.

Accelerating the Path to Novel Materials

Over 96% of manufactured goods, from food packaging to advanced technologies like semiconductors and solar panels, rely on chemicals that are difficult to replace. The discovery of alternative materials has traditionally relied on labor-intensive trial-and-error methods. Now, NVIDIA’s ALCHEMI microservice enables researchers to explore millions of chemical combinations in a virtual lab, dramatically reducing time, costs, and energy consumption.

Using AI-driven simulations, scientists can rapidly test material stability and identify promising compounds. For instance, running the MACE-MP-0 pretrained model on NVIDIA H100 Tensor Core GPUs achieves a 100x speedup in evaluating long-term stability. What once required months of computation can now be accomplished in hours, enabling researchers to examine 16 million structures with unprecedented efficiency.

Transforming Renewable Energy Research

By significantly accelerating chemical and material simulations, ALCHEMI is set to enhance research in renewable energy applications, such as the development of advanced solar panels and electric vehicle batteries. NVIDIA also plans to expand the platform with microservices that simulate the manufacturability of novel materials, helping scientists bring lab discoveries into practical use cases like fertilizers, pesticides, and sustainable products.

SES AI, a pioneer in lithium-metal battery development, is already leveraging the ALCHEMI microservice alongside the AIMNet2 model. This partnership enables SES AI to rapidly identify electrolyte materials for electric vehicles. According to SES AI CEO Qichao Hu, the tool has the potential to reduce discovery time and costs dramatically, accelerating innovation in battery technology. Recently, SES AI mapped 100,000 molecules in half a day using ALCHEMI and aims to scale this effort to analyze up to 10 billion molecules in the coming years.

Accessible Innovation for a Greener Future

The ALCHEMI microservice will soon be available for free testing via the NVIDIA NGC catalog and downloadable from build.nvidia.com. Researchers and developers will also have access to the production-grade NIM microservice through the NVIDIA AI Enterprise software platform, enabling seamless integration into existing workflows.

This initiative aligns with NVIDIA’s commitment to supporting the global Net Zero Initiative and empowering scientists, developers, and enterprises to meet energy and resource efficiency goals. With tools like ALCHEMI, NVIDIA is not only accelerating scientific discovery but also fostering sustainable innovation for a healthier planet.

Learn more about the NVIDIA ALCHEMI NIM microservice, and hear the latest on how AI and supercomputing are supercharging researchers and developers’ workflows by joining NVIDIA at SC24, running through Friday, Nov. 22.

See notice regarding software product information.