Roche Holding AG (OTC:RHHBY) is ramping up its artificial intelligence capabilities with a large-scale AI “factory” powered by advanced computing from NVIDIA Corp. (NASDAQ:NVDA), aiming to accelerate drug development timelines and improve diagnostics across its global operations.
Roche Expands AI Infrastructure Across US And Europe
The Swiss pharmaceutical giant on Monday said the new infrastructure includes 2,176 high-performance GPUs deployed on premises across the U.S. and Europe.
Combined with its cloud capabilities, Roche’s total AI computing capacity now exceeds 3,500 Blackwell GPUs, marking one of the largest AI footprints disclosed by a pharmaceutical company.
The initiative builds on a strategic collaboration with NVIDIA that began in 2023, reflecting Roche’s push to integrate AI across its full value chain — from early discovery to commercialization.
Executives Highlight Speed And Innovation Benefits
Wafaa Mamilli, Roche’s Chief Digital and Technology Officer, emphasized the role of speed in healthcare innovation.
“In healthcare, time is the most critical variable,” Mamilli said, noting that faster development cycles can bring treatments to patients sooner.
The AI factory combines large-scale computing with Roche’s scientific expertise to streamline processes across research, manufacturing, and diagnostics, she added.
Roche AI Factory Targets Drug Discovery And Manufacturing Efficiency
Roche said the platform will support its “Lab-in-the-Loop” approach, where AI models are integrated with real-world biology and chemistry experiments. Using NVIDIA’s BioNeMo platform, scientists can test hypotheses at scale and accelerate discovery timelines.
In manufacturing, the company is deploying digital twins — virtual replicas of production systems — using NVIDIA Omniverse tools to optimize factory design and operations.
The infrastructure also enhances diagnostics, with AI-powered tools analyzing large datasets and identifying subtle disease patterns through digital pathology systems.
Genentech R&D Head Sees Faster Path To Medicines
Aviv Regev, head of Genentech Research and Early Development, said the expanded computing power will enable more advanced predictive models.
Regev added that scaling Roche’s Lab-in-the-Loop strategy could significantly shorten the timeline from biological insights to life-saving treatments.
New Timeline For Residual Cancer Detection
In March, Droplet Biosciences, a diagnostics company developing lymph-based liquid biopsy tests, announced a collaboration with NVIDIA to detect residual cancer within 24 hours of surgery — weeks earlier than traditional blood-based monitoring.
Unlike conventional minimal residual disease (MRD) tests that rely on blood plasma weeks after surgery, Droplet collects lymphatic fluid just 24 hours after tumor removal.
Image Via Shutterstock
