Mayo Clinic has installed NVIDIA’s DGX SuperPOD infrastructure, featuring the new DGX B200 systems powered by the Blackwell architecture, to accelerate the development of foundation models for clinical use.

The Rochester, Minnesota-based health system is integrating the high-performance computing platform to support applications in digital pathology, precision medicine and drug discovery. According to Mayo Clinic, the system is already helping reduce the time required for AI model training, such as pathology slide analysis, from four weeks to one.

The deployment is part of Mayo’s broader digital transformation initiative under its "Bold. Forward." strategy, which includes advancing generative AI tools and foundation models to assist with clinical decision-making.
"Our aspiration for AI is to meaningfully improve patient outcomes by detecting disease early enough to intervene," said Dr. Matthew Callstrom, medical director of strategy and leader of Mayo’s Generative Artificial Intelligence Program. "What was once a hypothetical — 'If only we had the right data' — is now becoming reality, thanks to AI and advanced computing."

The system will initially focus on high-resolution imaging applications, supporting Mayo’s collaboration with Berlin-based Aignostics in developing Atlas, a pathology foundation model trained on over 1.2 million histopathology whole-slide images.

Jim Rogers, CEO of Mayo Clinic Digital Pathology, said the compute power is key to scaling Mayo’s clinical-grade models. “We’re transforming healthcare by quickly and safely developing innovative AI solutions that can improve patient outcomes and enable clinicians to dedicate more time to patient care,” Rogers said.

Mayo Clinic Digital Pathology oversees the institution’s AI-driven pathology initiatives, which include scanning, data storage, and model deployment. The group also works with startups and external partners to scale and commercialize its AI solutions.

Artificial Intelligence Homepage