NVIDIA’s AI infrastructure has significantly contributed to the National Artificial Intelligence Research Resource (NAIRR) pilot, helping to power over 700 U.S. research projects over the past two years. This initiative, led by the U.S. National Science Foundation, facilitates coordinated access to AI computational resources through public-private partnerships, supporting advancements in fields such as healthcare, agriculture, and energy. Utilizing NVIDIA’s DGX reference architecture, which provides dedicated GPU resources, researchers have been able to accelerate the development of groundbreaking technologies that tackle challenges in areas like physical sciences and infectious disease surveillance.

NVIDIA: NVIDIA develops AI hardware, software, and infrastructure solutions including the DGX platform. In June 2026, its cloud-based DGX resources and technical support powered the NAIRR pilot, enabling dedicated compute for U.S. research projects in fields like healthcare and energy. The company collaborates with academic teams to accelerate scientific AI workflows.
Polymathic AI: Polymathic AI is a coalition of scientists from institutions including Flatiron Institute, Cambridge University, and Lawrence Berkeley National Lab focused on building foundation models that leverage shared concepts across scientific disciplines. In recent work highlighted in June 2026, the group used NVIDIA GPUs and NVLink to develop and release the Walrus foundation model along with its Well dataset for fluidlike physical simulations.
Boston University: Boston University is a research university whose Hariri Institute for Computing and Center on Emerging Infectious Diseases advance computational approaches to public health challenges. In June 2026 reporting, its researchers leveraged NVIDIA accelerated compute through NAIRR to train and evaluate an LLM pipeline called BEACON for global infectious disease outbreak monitoring and analysis.
University of Michigan: University of Michigan is a major research university with expertise in engineering and computational sciences. In 2026, its team led by Professor Venkat Viswanathan used NAIRR allocations of NVIDIA DGX resources to develop the MIST family of models that fuse molecular AI with large language models for energy storage and materials discovery applications.

`json
{
“National Initiative”: “The NAIRR pilot provides researchers with coordinated access to AI computational resources through public-private partnerships.”,
“Research Focus Areas”: “NAIRR-supported projects address challenges in healthcare, agriculture, and energy using advanced AI models.”,
“Technology Enablement”: “NVIDIA DGX reference architectures deliver dedicated GPU resources that help researchers shorten development timelines for scientific foundation models.”
}
`