Bristol Myers Squibb deploys eight DGX Vera Rubin NVL72 systems for drug discovery AI factory

Bristol Myers Squibb is building what NVIDIA calls the life science industry's most advanced AI factory, deploying eight DGX Vera Rubin NVL72 systems as a second DGX SuperPOD the pharmaceutical company has nicknamed the "SuperDuperPOD."

What's new

According to NVIDIA's announcement, the new cluster is built on "eight DGX Vera Rubin NVL72 systems" and will "deliver up to 10x the performance per megawatt of the infrastructure it replaces." The system is designed for "running predictions, training models and powering agentic workflows across the full drug discovery pipeline" — spanning target identification, lead optimization, and molecule design.

Erin Davis, Bristol Myers Squibb's VP of Research Business Insights and Technology, described a shift from restricted to broad access: "Instead of equipping a small group of researchers with access to the supercomputer, we're opening it up to literally every scientist." Davis added that the company's existing infrastructure is already running at capacity: "We're saturated. We're in production with some very large-scale predictions around large molecules."

Use cases include AI-enabled target identification that Bristol Myers Squibb says "saves scientists weeks of manual work," an expanding library of CELMoD compounds for cancer-causing protein degradation research, and lead optimization using what the company calls a "Predict First" methodology. Payal Sheth, SVP of Therapeutic Discovery Sciences, framed the previous approach's limitation: "Every project was treated differently, and there were discrete sets of learnings that did not compound into any kind of intelligence framework."

Context

Bristol Myers Squibb already operates one of the largest AI compute clusters in the pharmaceutical sector, and this expansion builds directly on that existing DGX SuperPOD rather than starting from scratch. It lands alongside a wave of Vera Rubin-based deployments NVIDIA has announced across sectors this month, including at the Naval Postgraduate School, as the platform becomes NVIDIA's flagship offering for compute-intensive enterprise AI workloads.

Why it matters

Pharma R&D is one of the clearest cases where AI compute directly maps to institutional output — faster target identification and lead optimization can compress drug-discovery timelines that otherwise run years. Bristol Myers Squibb's move to open supercomputer access to every scientist, rather than a specialized team, is also a notable operating-model shift: it treats large-scale AI prediction as standard lab infrastructure rather than a scarce, gated resource, which is the kind of internal democratization other large pharma and biotech companies are likely to watch closely.