NSF and NVIDIA Partner to Launch Open-Source AI Models for Scientific Research

On August 14, 2025, the U.S. National Science Foundation (NSF) and NVIDIA announced a partnership to develop open-source artificial intelligence (AI) models aimed at advancing scientific research. The collaboration, known as the Open Multimodal AI Infrastructure to Accelerate Science (OMAI) project, is led by the Allen Institute for AI (Ai2). NSF will invest $75 million, while NVIDIA will contribute $77 million, totaling $152 million in funding. Read more

The OMAI project aims to create a suite of advanced, open-source, multimodal large language models trained on scientific data and literature. These models are designed to assist researchers in processing and analyzing research more efficiently, generating code and visualizations, and connecting new insights to past discoveries. The initiative also includes workforce development efforts to expand AI expertise beyond traditional tech hubs, thereby strengthening American competitiveness in AI and other critical technologies. Read more

NVIDIA's support includes providing HGX B300 systems—state-of-the-art AI infrastructure designed to accelerate model training and inference—with the NVIDIA AI Enterprise software platform. This infrastructure aims to empower OMAI to transform massive datasets into actionable intelligence and breakthrough innovations. Read more

Initial applications of Ai2's work will focus on accelerating the discovery of new materials, improving protein function prediction for biomedical advancements, and addressing core weaknesses in today's large language models. Read more

This partnership aligns with the White House AI Action Plan, which seeks to accelerate AI-enabled science and ensure the United States produces leading open models that enhance America's global AI dominance. Read more

Brian Stone, Acting NSF Director, emphasized the transformative impact of AI in scientific research, stating, "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." Read more

Jensen Huang, Founder and CEO of NVIDIA, highlighted the role of AI in modern science, saying, "AI is the engine of modern science — and large, open models for America's researchers will ignite the next industrial revolution." Read more

Ali Farhadi, CEO of Ai2, underscored the necessity of open AI ecosystems, stating, "Fully-open AI is not just a preference — it's a necessity. For the U.S. to continue leading the next era of scientific and technological discovery, we must create open, collaborative ecosystems." Read more

The OMAI project addresses the growing cost barrier of creating powerful AI models, which has exceeded the budgets of university labs and federally funded researchers. By providing open-source AI tools, the initiative democratizes access to advanced AI capabilities, enabling a broader range of researchers to contribute to scientific advancements. This approach fosters inclusivity and collaboration, potentially accelerating discoveries across various scientific fields. Read more

While there have been previous collaborations between public institutions and private companies to advance AI research, the scale and openness of the OMAI project are noteworthy. The commitment to fully open-source models, including training data, code, and documentation, sets a new standard for transparency and accessibility in AI research. This initiative builds upon prior efforts but distinguishes itself through its emphasis on openness and its substantial investment.

The comprehensive partnership between NSF, NVIDIA, and Ai2 represents a significant step toward integrating AI more deeply into scientific research, with the potential to accelerate discoveries and maintain the United States' leadership in technological innovation.