NVIDIA Hits $5T Valuation, Teams Up With Lilly To Build AI Supercomputer

Eli Lilly and Company is partnering with NVIDIA to build a supercomputer designed to handle data ingestion, model training, fine-tuning, and large-scale inference to accelerate drug discovery. The system will power an AI factory that enables scientists to identify, optimize, and validate new molecules while training AI models on millions of experiments to find potential medicines. It will also support manufacturing through digital twins and NVIDIA’s robotic technologies, improving production efficiency and reducing downtime. Additionally, the supercomputer will enhance medical imaging to better track disease progression and aid in developing biomarkers for personalized treatments.

The system will power enterprise AI agents to help researchers with reasoning, planning, and collaboration across digital and physical environments. According to Lilly, it is the world’s first NVIDIA DGX SuperPOD featuring DGX B300 systems, equipped with over 1,000 B300 GPUs connected by a unified high-speed network. The supercomputer will operate on 100% renewable energy within Lilly’s facilities and use the company’s chilled water infrastructure for liquid cooling.

Kimberly Powell, vice president of healthcare at NVIDIA, said the AI revolution will have its most profound impact on medicine, transforming how biology is understood. She described modern AI factories as new scientific instruments that enable a shift from trial-and-error discovery to intentional medicine design, positioning Lilly as a global leader in this new era.

The collaboration comes as NVIDIA became the world’s first company to reach a $5 trillion valuation after its stock surged more than 5%. The partnership follows broader industry efforts to build AI data centers for drug discovery, such as Project Stargate, a $500 billion initiative announced by President Donald Trump in collaboration with Oracle, OpenAI, and SoftBank to construct massive AI campuses across the U.S., starting in Abilene, Texas. NVIDIA also recently partnered with OpenAI to deploy at least 10 gigawatts of AI data centers using millions of its GPUs and invested $100 billion to support the project. Meanwhile, Lila Sciences has entered the race with its own AI-enabled scientific superintelligence computer, raising $550 million in under a year, including a $350 million Series A.