In a groundbreaking collaboration, NVIDIA has partnered with the U.S. National Science Foundation (NSF) to create an innovative AI system aimed at advancing scientific research across America.
This initiative supports the NSF’s Mid-Scale Research Infrastructure project known as the Open Multimodal AI Infrastructure to Accelerate Science (OMAI).
Brian Stone, who is currently performing the duties of the NSF director, acknowledged the transformative power of AI in scientific endeavors.
He stated, “Bringing AI into scientific research has been a game changer. 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.”
OMAI, an endeavor associated with the Allen Institute for AI (Ai2), aims to develop a fully open AI ecosystem that propels scientific discovery while simultaneously enhancing the science of AI itself.
NVIDIA’s commitment to OMAI includes the provision of its state-of-the-art HGX B300 systems—designed to streamline model training and inference with high efficiency.
These systems employ NVIDIA Blackwell Ultra GPUs and incorporate advanced high-bandwidth memory and interconnect technologies, delivering exceptional performance for the most demanding workloads.
“AI is the engine of modern science—and large, open models for America’s researchers will ignite the next industrial revolution,” said Jensen Huang, founder and CEO of NVIDIA.
He emphasized that the collaboration with NSF and Ai2 would facilitate innovative breakthroughs with top-tier infrastructure that enables U.S. scientists to harness boundless intelligence, thus transforming it into America’s most potent and renewable resource.
This initiative is set to benefit research teams from various institutions, including the University of Washington, the University of Hawaii at Hilo, the University of New Hampshire, and the University of New Mexico.
The partnership aligns with recent initiatives from the White House, outlined in the AI Action Plan, which aims to fortify America’s leadership in global AI technology.
Noah Smith, senior director of natural language processing research at Ai2, stressed the significance of NVIDIA’s support, stating, “The models are part of the national research infrastructure—but we can’t build the models without compute, and that’s why NVIDIA is so important to this project.”
Today’s large language models (LLMs), which have become integral to many rapidly growing applications, possess billions of parameters derived from extensive training datasets.
Multimodal LLMs expand this capability by integrating images, graphs, tables, and other forms of data.
However, the formidable potential of these frontier models often remains inaccessible for scientific research due to restricted availability of parameters, training data, code, and comprehensive documentation.
Smith elaborated, “With the model training data in hand, you have the opportunity to trace back to particular training instances similar to a response, and also more systematically study how emerging behaviors relate to the training data.”
The collaborative effort between NVIDIA and NSF to support Ai2’s OMAI project not only aims to provide unrestricted access to research models but also to offer open-source data interrogation tools and documentation.
This is particularly designed to aid early-career researchers and bolster the United States’ position in global science and engineering.
The Ai2 initiative, utilizing NVIDIA technologies, intends to make the software and models available at low or no cost, akin to free access provided by open-source code repositories and science-oriented digital libraries.
This strategy aligns with Ai2’s previous efforts to develop fully open language models and multimodal frameworks, thus ensuring maximum accessibility for researchers.
Further reinforcing the significance of this initiative, the ‘Winning the AI Race: America’s AI Action Plan’—announced in July by the White House—highlights steps taken to expedite federal permitting for data center infrastructure and promote the global export of American AI technologies.
OMAI’s foundations resonate with the priorities set by the White House AI Action Plan, focusing on propelling AI-enabled science and fostering the development of leading open models, which are crucial for enhancing America’s educational and research capabilities in the arena of artificial intelligence.
image source from：blogs
