Bristol Myers Squibb Expands Drug Discovery Computing with Vera Rubin

What happened: Bristol Myers Squibb announced the deployment of its second DGX SuperPOD, built on eight DGX Vera Rubin NVL72 systems, to expand the use of AI across drug research and development. According to NVIDIA, the new system delivers up to ten times the performance per megawatt of the infrastructure it replaces, and provides a unified platform for predictions, model training and agentic workflows using BioNeMo tools. The company says its existing system has already helped cut weeks of manual work from identifying drug targets.

The lens: economic transformation The investment links the value of computing to a measurable research outcome, not simply the number of accelerators. It also turns infrastructure from a resource confined to a specialist team into a shared layer that researchers can use across the company's sites.

Who is affected: Drug discovery researchers, computational chemistry and bioinformatics teams, infrastructure managers, and pharmaceutical companies weighing the purchase of computing capacity against building shared internal platforms.

What it means for the region: For research centres and pharmaceutical companies in the region, the most important signal is that returns depend on linking computing capacity to a specific discovery workflow and reusable data. Buying hardware without unifying access, data and validation tools may leave computing disconnected from laboratory decisions.

The practical takeaway: Select one costly stage of drug discovery, such as ranking candidate compounds, and establish a baseline for its duration and the quality of its decisions. Then test a limited prediction workflow with human review before estimating the need for broader infrastructure.

Source: https://blogs.nvidia.com/blog/bristol-myers-squibb-building-life-science-industrys-most-advanced-ai-factory-on-nvidia-vera-rubin/