Bristol Myers Squibb is expanding its AI drug discovery efforts by adding a second Nvidia DGX SuperPOD system. The new setup features eight DGX Vera Rubin NVL72 units, offering up to ten times better performance per megawatt compared to the previous generation hardware. This upgrade enables BMS to process heavier AI workloads efficiently across their research network.
Boosting AI-Driven Drug Discovery
This latest expansion builds on nearly three years of collaboration between Bristol Myers Squibb and Nvidia. The original DGX SuperPOD significantly reduced AI-enabled target identification times from weeks to just days. By integrating the two SuperPODs into a unified platform, BMS will remove previous access bottlenecks, allowing researchers worldwide to run AI tasks ranging from training proprietary foundation models to performing automated target validation with minimal human intervention.
Central to the strategy is BMS's “Predict First” approach. Rather than relying on traditional early-stage molecule synthesis and lab experiments, AI models predict promising candidates upfront. This process filters less viable options early, channeling resources toward the most promising drug programs. The enhanced system also expands BMS's capacity to explore larger chemical spaces and perform more sophisticated molecular predictions.
Software Integration and Therapeutic Coverage
Alongside the hardware upgrade, BMS gains access to Nvidia’s BioNeMo platform and the BioNeMo Agent Toolkit, optimized for biological and pharmaceutical AI workflows. These tools support research efforts across oncology, hematology, cardiovascular disease, immunology, and neuroscience, reflecting the broad scope of BMS’s development pipeline.
The companies have not disclosed financial details of this expansion. Despite this, BMS shares showed a modest increase of 0.31%, and Nvidia rose by 2.28% following the announcement.
material is for informational purposes only and not financial advice
