Lilly Launches World's Top Pharma AI Factory

🔑 Enhanced Key Takeaways

• •LillyPod was assembled in just four months and delivers over 9,000 petaflops of AI performance, capable of over 9 quintillion math problems per second[1][4].
• •The system runs on 100% renewable electricity within existing Lilly facilities and uses chilled water liquid cooling[7].
• •LillyPod supports applications beyond drug discovery, including medical imaging, manufacturing digital twins, enterprise AI agents, and biomarker discovery for gene therapies[1][7][8].

🛠️ Technical Deep Dive

• •Built with exactly 1,016 NVIDIA Blackwell Ultra GPUs (referred to as B300 systems)[1][4][7].
• •Powered by NVIDIA’s full-stack AI factory architecture, including accelerated computing, NVIDIA Spectrum-X Ethernet networking, and optimized AI software[1][6].
• •Managed by NVIDIA Mission Control software for orchestrating workloads, monitoring performance, and automating AI operations across more than 1,000 GPUs[1].
• •Utilizes NVIDIA BioNeMo platform to train AI foundation models combining millions of past experiments with public research for generating antibodies, nanobodies, and novel molecules[1].
• •Features a unified networking fabric for communication across GPUs, storage, and systems on a single high-speed network[7][8].

🔮 Future ImplicationsAI analysis grounded in cited sources

⏳ Timeline