Roche Expands AI Capabilities With NVIDIA-Powered Supercomputing To Accelerate Drug Development, Builds One Of Pharma’s Largest AI Infrastructures

Roche has announced a major expansion of its global AI infrastructure, deploying a large-scale AI factory powered by NVIDIA's advanced computing. The setup includes 2,176 high-performance GPUs installed across the United States and Europe, forming part of a broader system that now exceeds 3,500 Blackwell GPUs across on-premise and cloud environments, making it one of the largest AI computing footprints in the pharmaceutical industry.

This expansion builds on Roche’s strategic collaboration with NVIDIA, which began in 2023, and aims to accelerate innovation across the company’s entire value chain. By integrating cutting-edge AI infrastructure, Roche aims to enhance data-driven decision-making and accelerate the development of diagnostics and therapeutics.

“In healthcare, time is the most critical variable; every day saved means a life-changing medicine or diagnostic reaches a patient sooner,” said Wafaa Mamilli, Roche’s Chief Digital and Technology Officer. “Our AI factory combines world-class computing power with Roche’s scientific expertise to embed AI across the entire value chain, from discovery to development, manufacturing and commercialisation, transforming how we deliver the next generation of medicines and diagnostics solutions.”

At the core of this initiative is Roche’s AI factory, a high-performance supercomputing platform designed to drive digital transformation. In research and development, platforms like NVIDIA BioNeMo support Roche’s “Lab-in-the-Loop” approach, enabling seamless integration of experimental biology and chemistry with AI models. This allows scientists to test hypotheses at scale and unlock new scientific insights more efficiently.

Beyond R&D, the AI factory is also transforming manufacturing and diagnostics. Technologies such as digital twins, powered by NVIDIA Omniverse, help optimise production processes and facility design. In diagnostics and digital pathology, AI-driven tools analyse vast datasets and medical images to detect subtle disease patterns, while solutions like NVIDIA Parabricks accelerate genomic data processing. Additionally, Roche is using NVIDIA NeMo Guardrails to ensure safe and reliable deployment of conversational AI in healthcare settings.

“Our expanded collaboration with NVIDIA and the launch of this AI factory further strengthen our leadership in AI-driven drug discovery and development,” said Aviv Regev, Executive Vice President and Head of Genentech Research and Early Development (gRED). “By providing the massive computational power needed to continue to scale our Lab-in-the-Loop strategy, a space we have pioneered for over five years, our scientists can build more sophisticated predictive frontier models and further shorten the path from biological insight to life-saving medicine.”

Overall, this investment reflects Roche’s broader vision to embed AI across its operations, enhancing human expertise rather than replacing it. By expanding access to high-performance computing, the company aims to tackle complex disease challenges more effectively and accelerate progress toward preventing, diagnosing, and treating diseases worldwide.