AI Factory with NVIDIA Omniverse Digital Twins for GLP-1 Manufacturing

Roche launched one of the pharmaceutical industry's largest AI infrastructure deployments through a partnership with NVIDIA, standing up an AI factory anchored by more than 3,500 NVIDIA Blackwell GPUs (2,176 on-premises in the US and Europe, the remainder in cloud environments). The infrastructure powers Roche's full value chain from discovery to manufacturing. In drug discovery, the GPUs run a 'Lab-in-the-Loop' model that links automated biology and chemistry experiments with AI-driven hypothesis testing, compressing what previously took more than two years of iteration. The stack uses NVIDIA BioNeMo as the foundation-model platform for molecular design, Parabricks for genomic and pathology data analysis, and NeMo with NeMo Guardrails to enable safe enterprise-grade conversational AI for scientists. For manufacturing, Roche is using NVIDIA Omniverse to build high-fidelity digital twins of its new GLP-1 (incretin) plant in Holly Springs, North Carolina, so the facility can be designed, optimized, and de-risked virtually before being built. The site is Roche's Genentech unit's first East Coast facility, broken ground in August 2025 at $700M and expanded in early 2026 to $2.0B, dedicated to next-generation metabolic and obesity treatments including Carmot-derived GLP-1/GIP programs (CT-388, CT-868, oral CT-996). Industry analysis of digital-twin factories reports capacity increases of 25-40% and lead-time reductions of 15-20%; the goal is to compress drug-development cycles, optimize manufacturing throughput, and accelerate delivery of life-changing medicines across Roche's pharmaceuticals and diagnostics divisions. (Holly Springs, North Carolina, USA)