BMS deploys second NVIDIA DGX SuperPOD on 8 Vera Rubin NVL72 systems, delivering 10x perf/W for AI drug discovery.
Bristol Myers Squibb deployed its second NVIDIA DGX SuperPOD on eight Vera Rubin NVL72 systems. The cluster delivers up to 10x performance per megawatt and opens AI access to every researcher at the pharma giant.
Key facts
- 8 DGX Vera Rubin NVL72 systems deployed by BMS.
- 10x performance per megawatt vs prior infrastructure.
- AI saves weeks per target identification project.
- CELMoD compound library expanded via AI screening.
- BMS operates second SuperPOD after 3 years of first.
Erin Davis calls it the “SuperDuperPOD.” That’s two things in one name: pharmaceutical giant Bristol Myers Squibb (BMS) already runs one of the largest AI clusters in life sciences, with serious results to show for it. And they’re doubling down.
BMS announced today it is deploying its second NVIDIA DGX SuperPOD, this one built on eight DGX Vera Rubin NVL72 systems — the most powerful and energy-efficient AI cluster in life sciences. According to the NVIDIA blog, the eight rack-scale systems, each comprising NVIDIA Vera CPUs and Rubin GPUs, deliver up to 10x the performance per megawatt of the infrastructure it replaces. The new cluster will give researchers access to a unified AI platform including NVIDIA BioNeMo Agent Toolkit for biological AI, enabling predictions, model training, and agentic workflows across drug discovery.
“Instead of equipping a small group of researchers with access to the supercomputer, we’re opening it up to literally every scientist,” says Davis, vice president of research business insights and technology at BMS. “No one has to wait, and no one is told they have a limit.”
Measurable Impact from First SuperPOD
BMS has operated a DGX SuperPOD for about three years, producing meaningful results. AI-enabled target identification already saves scientists weeks of manual work, freeing time to focus on the highest-value scientific decisions. BMS’s team has used AI to expand its library of CELMoD compounds — molecules engineered to selectively degrade cancer-causing proteins, with applications in blood cancer treatment and beyond. This has opened the door to new targets and new potential medicines across a wider range of diseases.
AI is also applied in lead optimization stages of drug discovery using a methodology called “Predict First,” which informs experimental gating based on design predictions. “We use predictions as a way to prioritize synthesis of molecules with multi parameter optimization,” explains Payal Sheth, senior vice president of therapeutic discovery sciences at BMS, “to weed out molecules that wouldn’t necessarily meet the property landscape we’re working towards. This ensures precious laboratory experiments are aligned with progressing molecules that have the highest probability of success.”
These research AI applications have significant impact on compute needs across the research organization. “We’re saturated,” Davis says. “We’re in production with some very large-scale predictions around large molecules. We’re building our” — the new cluster, the SuperDuperPOD.
Why This Matters: Democratizing Compute
The mandate, says Sheth — a scientist who spent her career inside drug discovery labs before taking on an expanded role in January as senior vice president of therapeutic discovery sciences at BMS — is moving from “sort of this abstract position of what AI can do to actually translating that to measurable impact.” The new cluster represents a structural shift: instead of a few researchers queueing for GPU time, every BMS scientist gets access. That could accelerate the pipeline from target identification to lead optimization, compressing years of wet-lab work into weeks of AI-driven prediction.
Nvidia’s Vera Rubin platform, announced in mid-2026, is the company’s next-generation AI infrastructure, featuring liquid-cooled designs and significant compute density. Nvidia CEO Jensen Huang pledged delivery of “giant amounts” of Vera Rubin chips in July 2026. [Per the source], the Vera Rubin NVL72 cloud rollout is expanding to Europe. The BMS deployment is among the first large-scale life science applications of the new hardware.
What to Watch
Watch for BMS’s first disclosure of how many molecules the new cluster has prioritized in lead optimization, and whether the CELMoD compound library expands to cover new cancer targets within the next two quarters. Also track Nvidia’s Vera Rubin supply: if BMS’s “SuperDuperPOD” scales to additional sites, it signals Vera Rubin is meeting demand despite previous manufacturing delays.
Source: blogs.nvidia.com
[Updated 22 Jul via gn_gpu_cluster]
New benchmark data [per Wccftech] reveals that Vera Rubin NVL72 achieves 800,000 tokens/s at 150MW, a 10x uplift over Blackwell’s 80,000 tokens/s at the same power envelope — underscoring the raw throughput BMS gains with its eight-system cluster.
Originally published on gentic.news
Top comments (0)