Lilly Launches World’s Most Powerful AI Factory for Drug Discovery Using NVIDIA Blackwell
Indianapolis — Eli Lilly and Company this week activated LillyPod, the pharmaceutical industry’s most powerful wholly owned and operated AI factory, built on NVIDIA’s latest DGX SuperPOD technology. The system, powered by 1,016 NVIDIA Blackwell Ultra GPUs, is designed to accelerate drug discovery, genomics research and personalized medicine development at unprecedented scale and speed.
Announced at NVIDIA GTC Washington, D.C., LillyPod represents the first deployment of an NVIDIA DGX SuperPOD featuring DGX B300 systems by a pharmaceutical company. Lilly states the new AI infrastructure will enable its researchers to compress traditional drug discovery timelines, improve accuracy of molecular modeling, and pursue breakthroughs that were previously computationally infeasible.
According to NVIDIA’s official blog, the system is “the world’s most powerful AI factory wholly owned and operated by a pharmaceutical company.” It builds on a broader collaboration between the two companies, including a previously announced AI co-innovation lab unveiled at the J.P. Morgan Healthcare Conference focused on tackling persistent challenges in pharmaceutical R&D through advanced AI.
Technical Scale and Capabilities
LillyPod leverages 1,016 NVIDIA Blackwell Ultra GPUs organized in a DGX SuperPOD configuration. This architecture delivers massive parallel computing power optimized for large-scale AI training and inference workloads common in modern drug discovery pipelines, including protein structure prediction, molecular dynamics simulations, and generative AI models for compound design.
The deployment marks a significant milestone in the convergence of AI and life sciences. Traditional drug discovery often requires years of iterative laboratory testing. By bringing this level of GPU-accelerated computing in-house, Lilly aims to shift more of the discovery process into the computational domain, potentially reducing both time and cost while increasing the probability of clinical success.
NVIDIA and Lilly have been deepening their partnership over the past year. The companies’ joint AI co-innovation lab was established to develop new AI methodologies specifically tailored to pharmaceutical challenges, ranging from target identification to clinical trial optimization.
Strategic Importance for Pharma
For Eli Lilly, LillyPod is more than an infrastructure upgrade — it is a strategic bet that AI will become the defining competitive advantage in next-generation drug development. The company has aggressively invested in AI and machine learning across its research organization in recent years, particularly in areas such as diabetes, obesity, and oncology.
The scale of LillyPod — over one thousand latest-generation Blackwell Ultra GPUs — places it among the most powerful single-company AI systems deployed in the life sciences sector. Most pharmaceutical companies currently rely on public cloud providers or smaller on-premise clusters for AI workloads. Lilly’s decision to build and operate its own SuperPOD signals a desire for complete control over data sovereignty, model training, and research IP.
NVIDIA, for its part, continues to expand its footprint in healthcare and life sciences. The company has positioned its DGX and Grace-Blackwell platforms as foundational infrastructure for what it calls “AI factories” — specialized supercomputing environments dedicated to industrial AI applications.
Impact on Developers, Researchers and the Industry
For researchers at Lilly, the new system means access to vastly greater computational resources directly integrated into their workflows. This could accelerate everything from training large biology foundation models to running high-throughput virtual screening of billions of potential drug candidates.
For the broader pharmaceutical industry, LillyPod sets a new benchmark. Other large pharma companies are expected to evaluate similar large-scale AI infrastructure investments as the cost of training state-of-the-art scientific AI models continues to grow. The deployment also validates NVIDIA’s strategy of offering turnkey DGX SuperPOD systems tailored for enterprise AI rather than relying solely on hyperscale cloud providers.
For AI developers and startups working in computational biology, the news reinforces the increasing importance of massive GPU clusters for cutting-edge research. It may also create new opportunities for collaboration as Lilly looks to expand its AI talent pool and partner ecosystem.
What’s Next
Lilly has not publicly disclosed detailed performance benchmarks for LillyPod or specific drug programs that will immediately leverage the system. The company is expected to share early results and use cases in the coming months as the AI factory moves into full production.
The activation of LillyPod comes amid rapid advancement in both AI hardware and scientific AI models. NVIDIA’s Blackwell Ultra platform itself is still rolling out across global data centers, making Lilly one of the earliest adopters at this scale in the pharmaceutical sector.
Industry observers will be watching whether LillyPod delivers measurable reductions in drug discovery timelines and improvements in clinical success rates — outcomes that could determine whether other major pharmaceutical companies follow suit with similar billion-dollar-scale AI infrastructure builds.
As the partnership between NVIDIA and Lilly matures through their co-innovation lab, further joint developments in specialized AI models for biology and chemistry are anticipated. Both companies have described the collaboration as a “blueprint for what is possible” when state-of-the-art AI compute meets deep pharmaceutical domain expertise.
This article is based on official announcements from NVIDIA and Eli Lilly and Company.