US scientific research groups receive $150M investment from NVIDIA and the National Science Foundation to develop shared multimodal AI models.
The U.S. National Science Foundation (NSF) and NVIDIA have announced a partnership worth over $150 million to develop open, multimodal AI models that will revolutionize scientific research and discovery. This initiative, led by the Allen Institute for AI (Ai2) and called the Open Multimodal AI Infrastructure to Accelerate Science (OMAI), focuses on speeding breakthroughs in various fields such as materials science, biology, and energy.
The partnership aims to equip America's scientists with tools to accelerate breakthroughs, secure U.S. global leadership in science and technology, and tackle challenges once thought impossible. The resources include NVIDIA HGX B300 systems built with Blackwell Ultra GPUs and the NVIDIA AI Enterprise software platform.
This five-year project, with a joint funding of $152 million from Nvidia and NSF, positions Ai2 as a central player in building AI infrastructure for science in the U.S. The AI models will be fully open and multimodal, capable of understanding and integrating multiple types of scientific data and literature, which is intended to supercharge research productivity and innovation across disciplines.
The project leader, Noah Smith, emphasizes AI not just as a tool but a catalyst for rethinking the process of scientific discovery itself, accelerating how new insights and technologies come to light. The initiative marks NSF’s first major investment in AI infrastructure for science, emphasizing open and shared AI resources to serve the scientific community nationwide.
The models can process text, images, graphs, and tables, allowing researchers to analyze findings faster and connect new insights to past discoveries. Potential applications of these models include accelerating the discovery of new materials, improving protein function prediction, and addressing weaknesses in current large language models.
Ali Farhadi, CEO of Ai2, stated that fully-open AI is a necessity for the U.S. to continue leading the next era of scientific and technological discovery. The partnership will support research teams at several universities, including the University of Washington, the University of Hawaii at Hilo, the University of New Hampshire, and the University of New Mexico.
This large-scale investment and infrastructure development signal a strategic push to integrate advanced AI deeply into the U.S. scientific ecosystem, enabling faster, more efficient discoveries with potentially transformative impacts on technology and science sectors.
- The partnership between the National Science Foundation (NSF) and NVIDIA, worth over $150 million, aims to utilize cutting-edge AI models to expedite breakthroughs in various scientific fields, such as materials science, biology, and energy, with a focus on U.S. global leadership in technology and science.
- To accelerate scientific discovery, the Open Multimodal AI Infrastructure to Accelerate Science (OMAI) initiative will invest in AI infrastructure, providing tools like NVIDIA HGX B300 systems and software to researchers across America.
- The AI models derived from this project, capable of understanding and integrating multiple types of scientific data, will facilitate faster research processes, allowing researchers to analyze findings more efficiently, connecting new insights to past discoveries, and ultimately driving innovation in both technology and science.
