Bristol Myers Squibb Expands AI Drug Discovery With NVIDIA Vera Rubin SuperPOD

Bristol Myers Squibb (BMS) is expanding its artificial intelligence (AI) infrastructure with a new NVIDIA DGX SuperPOD built on the company's Vera Rubin architecture, becoming the first life sciences company to deploy the latest-generation system for drug discovery and development.

The pharmaceutical company said the new deployment will support AI across its research organization, allowing scientists to train larger foundation models, run large-scale biological predictions, and use agentic AI workflows throughout the drug discovery pipeline. According to NVIDIA, the system consists of eight DGX Vera Rubin NVL72 systems and delivers up to 10x more performance per megawatt than the infrastructure it replaces.

The new AI cluster will operate alongside BMS' existing DGX SuperPOD, creating a unified computing environment accessible across the company's global research organization. The platform will also incorporate NVIDIA BioNeMo Agent Toolkit and NVIDIA Mission Control to support biological AI models and AI infrastructure management.

The investment reflects a broader shift toward dedicated AI infrastructure across enterprise research organizations, where compute capacity is becoming as important as proprietary data and AI models.

"Our mandate is moving from sort of this abstract position of what AI can do to actually translating that to measurable impact," Payal Sheth, Senior Vice President of Therapeutic Discovery Sciences at Bristol Myers Squibb, said in NVIDIA's announcement.

AI Compute Becomes a Strategic Asset for Pharmaceutical R&D

Reuters reported that financial terms of the purchase were not disclosed. The company said it previously deployed a smaller DGX SuperPOD that is now several generations behind the Vera Rubin platform. 

Robert Plenge, Chief Research Officer at Bristol Myers Squibb, told Reuters the additional computing power will allow researchers to evaluate significantly more potential drug candidates early in the discovery process. He also said the company is already using AI tools to reduce the time required to produce medicines for clinical testing by 20% to 30%, with that improvement potentially reaching 50% over time. Plenge added that one experimental sickle cell disease treatment currently in clinical development likely would not have been discovered without AI-enabled research.

Greg Meyers, Chief Digital and Technology Officer at Bristol Myers Squibb, said growing demand for larger AI models across research programs drove the investment. According to Reuters, AI is now used across all of the company's small-molecule drug programs and most of its large-molecule programs. Meyers also cited the system's improved energy efficiency, noting that delivering substantially more compute per watt is becoming increasingly important as organizations scale AI infrastructure.

The deployment illustrates how pharmaceutical companies are moving beyond pilot AI projects toward purpose-built computing infrastructure that supports discovery at enterprise scale. By expanding compute capacity alongside proprietary data and internal AI models, Bristol Myers Squibb aims to accelerate scientific research while enabling researchers across the organization to access advanced AI capabilities without traditional infrastructure bottlenecks.