Bristol Myers Squibb (BMS) is scaling its NVIDIA-powered AI infrastructure to create what it describes as the ‘most powerful AI factory in life sciences’. A second system will deliver up to ten times greater performance per megawatt than its predecessor, deploying an NVIDIA DGX SuperPOD with DGX Vera Rubin NVL72 systems to create the life sciences industry’s most powerful and energy-efficient privately owned NVIDIA infrastructure. Greg Meyers, Chief Digital and Technology Officer, stated BMS has made a deliberate bet on AI and is beginning to see it pay off in its pipeline and operations. BMS built its first NVIDIA-powered DGX SuperPOD three years ago. AI agents that automate target identification and validation save researchers weeks of manual work. The company’s ‘Predict First’ approach uses AI to guide experimental design before laboratory testing, informing every small-molecule programme and most large-molecule programmes. Robert Plenge, Executive VP and Chief Research Officer, said the infrastructure lets scientists learn from every experiment and clinical readout to sharpen the next hypothesis. The expanded agreement supports a research vision of ‘hybrid intelligence’, where AI co-scientists work alongside human researchers. The cluster supports foundation models trained on decades of proprietary data and facilitates AI-driven workflows.