aiPublished on July 20, 20266 min read

Bristol Myers Squibb Builds the Most Advanced AI Factory in the Pharmaceutical Industry with NVIDIA Vera Rubin

Bristol Myers Squibb is deploying its second NVIDIA DGX SuperPOD, built on eight Vera Rubin NVL72 systems, to accelerate drug discovery with agentic AI.

Inteligência ArtificialNVIDIAAI InfrastructureHealthcareCiências da VidaIA AgenteSupercomputaçãoBusiness Automation

Bitclever AI Research

Author: Bitclever AI Research ## Executive Summary Pharmaceutical company Bristol Myers Squibb (BMS) has announced the deployment of its second NVIDIA DGX SuperPOD, this time built on eight DGX Vera Rubin NVL72 systems — considered the most powerful and energy-efficient AI cluster currently serving the life sciences industry. This investment reinforces BMS's commitment to democratising access to AI supercomputing for its entire research community, with the goal of measurably accelerating the drug discovery pipeline. ## What Happened Bristol Myers Squibb announced the deployment of a new NVIDIA DGX SuperPOD, internally nicknamed the "SuperDuperPOD" by Erin Davis, vice president of research business insights and technology at the company. This new system comprises eight DGX Vera Rubin NVL72 units, each integrating NVIDIA Vera CPUs and Rubin GPUs, and represents the second infrastructure of its kind installed by the pharmaceutical company, which has already been operating a DGX SuperPOD for around three years with proven results. According to NVIDIA, this new configuration delivers up to 10 times more performance per megawatt compared to the infrastructure it replaces, positioning it as the most advanced and energy-efficient AI cluster in the life sciences sector to date. The new infrastructure will support a unified AI platform, which includes the NVIDIA BioNeMo Agent Toolkit — a tool dedicated to biological AI — allowing BMS researchers to run predictions, train models, and power agentic workflows across the entire drug discovery pipeline. According to Erin Davis, the strategic goal is clear: rather than reserving access to the supercomputer for a restricted group of researchers, BMS is opening it up to all scientists in the organisation. "No one has to wait, and no one is told they've hit a limit," says Davis. Payal Sheth, who took on the role of senior vice president of therapeutic discovery sciences in January after a career dedicated to drug discovery labs, sums up the company's mandate: moving from an abstract stance on AI's potential to concretely translating that potential into measurable impact. ## Why This Matters This announcement comes at a time when the pharmaceutical industry faces growing pressure to reduce the time and costs associated with discovering new drugs, a traditionally slow and expensive process. The adoption of large-scale AI infrastructure, such as the Vera Rubin-based DGX SuperPOD, signals a paradigm shift: high-performance computing is no longer a scarce resource reserved for specific teams, but is becoming a capability that spans the entire scientific organisation. The fact that BMS has already been operating an AI cluster for three years, with documented results, reinforces the thesis that this type of investment is not merely experimental, but a central component of the company's research strategy. The introduction of tools like the NVIDIA BioNeMo Agent Toolkit also highlights the growing relevance of agentic AI — systems capable of executing complex workflows more autonomously — in scientific and biological research contexts. For the life sciences sector as a whole, this move sets a new competitive benchmark: organisations that fail to keep pace with this level of investment in AI infrastructure risk falling behind in terms of innovation speed and the ability to explore broader chemical spaces. ## Business Impact While this case centres on the pharmaceutical industry, the implications extend to any organisation that relies on data-intensive research and complex computational modelling. Some key points for business decision-makers include: - **Democratising access to high-performance computing**: BMS's decision to open supercomputer access to all researchers, rather than just a technical elite, illustrates how infrastructure efficiency can translate into broader organisational productivity gains. - **Energy efficiency as a strategic factor**: the up to 10x gain in performance per megawatt is not just a technical metric — it has direct implications for operating costs and sustainability, factors increasingly scrutinised by investors and regulators. - **Agentic AI applied to scientific workflows**: the integration of tools like the BioNeMo Agent Toolkit demonstrates how agentic AI is evolving from an experimental concept into an operational component of research and development processes. - **The need to align infrastructure with business strategy**: as Payal Sheth emphasises, the real challenge is not just having access to cutting-edge technology, but ensuring that access translates into measurable impact on business objectives. For companies in other sectors, the BMS case serves as a reference for how to structure AI infrastructure investments to maximise organisational return, rather than concentrating resources on isolated initiatives. ## Bitclever Perspective At Bitclever, we closely follow cases like Bristol Myers Squibb's because they illustrate a trend we've already observed among clients across various sectors: the transition from AI pilot projects to business-critical infrastructure, integrated across the entire organisation. BMS's experience reinforces three principles we apply when guiding our clients in defining AI and automation strategies: 1. **Infrastructure should serve strategy, not the other way around.** Investment in computing capacity only generates value when it's aligned with clear, measurable business objectives — as Payal Sheth emphasises in her approach to AI at BMS. 2. **Democratising access to technology is a value multiplier.** Enabling more employees to use advanced AI tools, without artificial access barriers, tends to accelerate innovation more consistently than concentrating resources within a restricted team. 3. **Agentic AI is becoming operational, not just experimental.** Tools like the BioNeMo Agent Toolkit show how agentic workflows are already being applied in highly complex scientific contexts — a trend we also observe in business process automation in Portugal and across the European market. Bitclever helps organisations assess where and how AI and automation investments can generate real impact, adjusting the scale of investment to the maturity and specific needs of each business, without needing to replicate infrastructure on the scale of pharmaceutical giants. ## Conclusion Bristol Myers Squibb's investment in a second DGX SuperPOD based on NVIDIA Vera Rubin represents more than a technology upgrade: it's a clear signal that large-scale AI is becoming critical infrastructure in the pharmaceutical industry. For organisations across all sectors, the key lesson is that the true value of AI lies not just in the available computing capacity, but in how that capacity is distributed, integrated, and aligned with measurable business objectives. As more companies follow this path, the ability to translate technology investment into concrete impact will be the decisive factor for competitive differentiation.

Sources

https://blogs.nvidia.com/blog/bristol-myers-squibb-building-life-science-industrys-most-advanced-ai-factory-on-nvidia-vera-rubin/

← Back to Blog

Bristol Myers Squibb Builds the Most Advanced AI Factory in the Pharmaceutical Industry with NVIDIA Vera Rubin

Sources

Related Articles

NVIDIA at SIGGRAPH 2026: Agentic AI and Physics Redefine Graphics and Simulation

Apple's Lawsuit Against OpenAI: What Impact on Hardware Plans and IPO?

Apple Sues OpenAI Over Trade Secrets: A Case That Could Jeopardise the Company's IPO