NVIDIA, NSF join forces with nonprofit to bring AI to scientific research

A new partnership between chipmaker NVIDIA, nonprofit research center the Allen Institute for Artificial Intelligence, and the National Science Foundation will pursue new development of multimodal AI models dedicated to scientific research.

Announced on Thursday, the Open Multimodal AI Infrastructure to Accelerate Science is a joint effort to further the U.S. AI infrastructure for scientific domains. Multimodal large language models are AI models that can process multiple different types of data, a flexibility that allows them to generate advanced outputs.

Trained on diverse data, a given multimodal AI model could generate advanced scientific outputs for researchers to further their work in arenas like biology, energy and materials sciences. The goal of the initiative is to ensure these models are designed to be open and available to the broader scientific community. 

“Bringing AI into scientific research has been a game changer,” said NSF Chief of Staff and acting Director Brian Stone. “NSF is proud to partner with NVIDIA to equip America’s scientists with the tools to accelerate breakthroughs. These investments are not just about enabling innovation; they are about securing U.S. global leadership in science and technology and tackling challenges once thought impossible.”

The Allen Institute for Artificial Intelligence, or Ai2, will lend their expertise to create multimodal large language models trained on specific scientific data and literature. NVIDIA will provide hardware and software, namely its HGX B300 systems and AI Enterprise software platform, in addition to supplying $77 million to the project.

The initial applications of the OMAI program will focus on biomedical research, specifically within materials discovery and protein function prediction, as well as addressing major faults in large language models. 

“AI is the engine of modern science — and large, open models for America’s researchers will ignite the next industrial revolution,” Jensen Huang, NVIDIA CEO, said in a statement. “In collaboration with NSF and Ai2, we’re accelerating innovation with state-of-the-art infrastructure that empowers U.S. scientists to generate limitless intelligence, making it America’s most powerful and renewable resource.”

NSF will also be contributing $75 million to the OMAI Initiative. This support is provided via the Mid-scale Research Infrastructure program, a mission to spearhead scientific research infrastructure by funding “high-impact, high-reward” research projects at a local and national level.

This announcement speaks to several priorities in the Trump administration’s AI policy strategy, such as the growth of U.S. AI infrastructure, more public-private partnerships and keeping leading AI models open-source. 

The latter policy position is a stance that Ali Farhadi, CEO of Ai2, said is a “necessity” for the U.S. to lead in the AI-powered scientific and technological race. 

“For the U.S. to continue leading the next era of scientific and technological discovery, we must create open, collaborative ecosystems where millions of researchers and developers can work together to improve and expand these systems,” Farhadi said. “This infusion will supercharge the work we do at Ai2 and increase America's ability to deliver breakthrough AI developments.”