Overview
- Bristol Myers Squibb announced Monday, July 20, 2026, that it is acquiring a second DGX SuperPOD built from eight DGX Vera Rubin NVL72 systems to expand AI across its drug discovery and development operations.
- The new cluster will be combined with BMS’s existing SuperPOD into a single data plane that researchers at all BMS sites can access for model training, predictions, and agentic workflows.
- NVIDIA says the Vera Rubin architecture can deliver up to 10 times the performance per megawatt versus the prior generation and BMS plans to run NVIDIA’s BioNeMo platform and the BioNeMo Agent Toolkit on the system.
- BMS says its earlier SuperPOD already cut target‑identification work from weeks to days and reduced computing costs by roughly 55%, and the new deployment is aimed at scaling those gains across small molecules, biologics, clinical applications, and digital twins.
- Financial terms were not disclosed, and the move follows a wider pharma trend of building dedicated, colocation‑hosted GPU supercomputers with partners such as Equinix and systems integrators to speed research and lower operating costs.