A first-of-its-kind AI lab from Lilly and NVIDIA aims to create continuous, data-driven systems that accelerate drug discovery.
NVIDIA and Eli Lilly have announced a first-of-its-kind AI co-innovation lab aimed at rebuilding the drug discovery process from the ground up. The two companies will invest up to $1 billion over five years to bring together advanced AI modeling, large-scale data generation, robotics and automated agentic wet-lab systems.
The lab will operate in the San Francisco Bay Area, where Lilly researchers and NVIDIA AI engineers will work side by side to create continuous learning systems that link real-world experiments with computational models. The goal is to accelerate identification, optimization and validation of potential medicines, with NVIDIA’s BioNeMo platform and next-generation architectures such as Vera Rubin at the core of the effort.
The collaboration builds on Lilly’s existing AI supercomputer and expands opportunities to use AI across clinical development, manufacturing and supply chain reliability. Both companies view this as a blueprint for the next era of drug discovery: rapid experimentation, scalable data generation and AI models custom-built to support scientific decision making.
Why This Matters for Data Analysts
The Lilly–NVIDIA collaboration shows how quickly AI-assisted workflows are becoming central to scientific work. Several themes are directly relevant to analysts who work with automation and model-driven processes:
- Continuous learning systems that operate much like live data pipelines.
- Agentic workflows that mirror automated validation, scheduling and model retraining.
- Domain-specific foundation models built with the same principles analysts use for fine-tuning and deployment.
- High-quality, purpose-built datasets that improve model accuracy and speed.
- Human-in-the-loop design that aligns with analyst-in-the-loop decision making.
These developments point to a future where analysts work with AI systems that ingest data, refine models and prepare the next set of actions automatically, while human judgment guides direction and interprets results.