BMS needed a shared AI environment that could help scientists reason over huge biological, chemical, imaging, and clinical datasets across global R&D, because isolated compute and manual analysis slowed target discovery, lead optimization, imaging research, and clinical insight generation.
Reported impact
BMS reports 55 percent overall cost savings versus its prior computing model, weeks of manual work eliminated in target identification, new targets and potential medicines opened through AI-assisted CELMoD expansion, and imaging and clinical models that give scientists higher-conviction signals faster across a globally shared AI platform.
BMS established an AI factory on NVIDIA DGX SuperPOD to unify computational science across target identification, molecular design, imaging, and clinical development. The platform connects research sites and lets scientists shift resources among large language model training, deep learning, imaging, and inference workloads on one hybrid environment managed with Equinix and Mark III Systems. BMS uses the platform to expand its CELMoD compound library, apply its Predict First methodology to lead optimization, and integrate the NVIDIA BioNeMo Agent Toolkit into broader biological and agentic AI workflows. In imaging, teams process CT and MRI data with NVIDIA MONAI and DGX SuperPOD, masking 80–90 percent of scans during self-supervised learning to build foundation models that match human lesion and tumor detection while also quantifying muscle, fat, and bone density. BMS is also applying transformers to clinical-trial data and expanding the environment with a second DGX SuperPOD powered by DGX Vera Rubin NVL72 systems to make the platform globally accessible across R&D.
Biological, chemical, imaging, and clinical-trial data flow into the centralized AI factory, DGX SuperPOD resources train or fine-tune models for target identification, molecule design, imaging, and patient-response prediction, and the outputs are fed back into scientist decisions and globally shared R&D programs.
55% overall cost savings compared with BMS’s prior computing model
AI-enabled target identification saves scientists weeks of manual work.
BMS plans to combine its existing and expanded resources into a unified environment with a common data plane accessible from every BMS site.
The team used NVIDIA MONAI to process and transform the data, then trained the model on DGX SuperPOD to automatically segment lesions, tumors, and organs in standard CT scans.
Training combined internal clinical data, public datasets, and scientific literature. Using DGX SuperPOD, BMS models outperformed standard baseline transformers.
Review how this reference was scored and trace each normalized claim back to its documented source.
Scoring methodology
Customer Reference Value measures usefulness as a customer reference. Evidence Confidence measures the quality and traceability of available evidence, not independent verification.
Traceability
Field coverage shows which facts each source supports. Dates are shown only when documented by the source.
Related
NVIDIA Sample Sep 22, 2026 DELFI Diagnostics DELFI Diagnostics needed to speed genomic analysis for cancer screening and therapeutic monitoring, but CPU-bound alignment and somatic variant-calling steps created turnaround-time and cost bottlenecks that limited how… 10–20 business days DeepSomatic · MutectCaller · NVIDIA Parabricks NVIDIA Sample Jul 10, 2026 Astellas Pharma Inc. Astellas needed to shorten the slow handoff between experimental scientists and computational chemists in early drug discovery so more researchers could use structure prediction and affinity screening without waiting on… More than 50 regular users and over 2,000 total users within four months NVIDIA BioNeMo · NVIDIA Boltz-2 NIM · NVIDIA DGX H100/H200 NVIDIA Sample Apr 23, 2026 Biofy Hospitals needed pathogen identification and antibiotic-resistance analysis much faster than five-day culture and antibiogram workflows, but CPU-bound genomic processing and limited AI expertise made it hard to turn… 100,000+ synthetic DNA sequences generated by the biology model Abby · H100 Tensor Core GPUs · NVIDIA CUDA-accelerated libraries 