Bristol Myers Squibb bets big on Nvidia AI to speed up drug discovery
Bristol Myers Squibb (BMS) is making a bold move to accelerate drug discovery by purchasing the first Nvidia DGX SuperPOD built on the company’s cutting-edge Vera Rubin architecture. The new system will integrate seamlessly with BMS’s existing AI infrastructure, doubling down on a strategy that has already reshaped how the pharmaceutical giant approaches research.
A leap in computational power
The deal centers on eight DGX Vera Rubin NVL72 systems, each combining Nvidia’s latest central processing units and graphics processing units. This hardware leap is designed to handle the growing demands of BMS’s AI models, which are now larger and more complex than before. The system will process compounds, proteins, and vast datasets to train proprietary models and run predictive analytics across research programs. While financial terms remain undisclosed, BMS executives emphasize that the Vera Rubin SuperPOD is at least two generations ahead of their current setup—a DGX SuperPOD they’ve relied on for about three years.
Breaking bottlenecks in research
BMS’s existing infrastructure is already operating at full capacity, strained by the need to screen large molecules and develop internal foundation models. The new system aims to eliminate the bottlenecks that have forced researchers to wait for access or accept limits on computational power. “Before, we could evaluate 10 potential candidates; now we can assess dozens,” said Robert Plenge, BMS’s chief research officer. The company’s “Predict First” approach uses AI predictions to filter out unsuitable molecules early, reducing manual research time and focusing lab resources on the most promising compounds.
A shift toward accessible, scalable AI
Greg Meyers, BMS’s chief digital and technology officer, highlights the growing need for scalable computing as AI models become more integral to research. The expanded environment will integrate both the new Vera Rubin system and the older SuperPOD into a shared global computing network, giving more scientists direct access to high-performance resources. Erin Davis, vice president of research business insights and technology, notes that the demand stems from large-scale predictions and the development of internal models that require significant graphics processing capacity.
Why it matters
This investment signals a turning point for AI in drug discovery, where computational bottlenecks often delay breakthroughs. By expanding access to advanced AI tools, BMS is positioning itself to evaluate more candidates earlier, potentially cutting development timelines. The move also underscores how specialized AI hardware is becoming indispensable for pharmaceutical innovation, setting a benchmark for competitors in the race to harness artificial intelligence for medical advancements.
Source: AI News. AI-assisted editorial synthesis — TechnoExpress.