Lilly and NVIDIA announce a drug discovery supercomputer

Eli Lilly and NVIDIA announced on October 28, 2025, that they are building what the companies call the most powerful supercomputer owned and operated by a pharmaceutical company ^[1]. The system will use more than 1,000 NVIDIA B300 GPUs and is the first NVIDIA DGX SuperPOD built with DGX B300 systems, all connected on a single high-speed network ^[1]. Lilly CEO Dave Ricks later said the machine was built in Indianapolis, where the company is headquartered ^[2].

Lilly describes the project as an "AI factory," a computing setup meant to handle everything from gathering data to training AI models to running them at high volume ^[1]. According to the company, scientists will use it to train AI models on millions of experiments testing potential medicines, which Lilly says will expand the scope of its drug discovery work ^[1]. Some of these AI models will be shared through Lilly TuneLab, a platform the company says is meant to give other researchers in the biopharma industry access to advanced discovery tools ^[1].

Lilly says the supercomputer will also be used beyond early drug discovery. The company plans to apply it to manufacturing, using digital twins and robotics to try to improve production and cut downtime, and to medical imaging, where it says AI could help track how diseases progress and support new biomarkers ^[1]. Diogo Rau, Lilly's chief information and digital officer, said the company believes no other firm in its industry is investing in AI infrastructure at this scale, calling the company's decades of data "one of our most powerful assets" ^[1]. Lilly also said the system will run on 100% renewable electricity within existing Lilly facilities, consistent with the company's goal of carbon neutrality by 2030 ^[1].

At a separate event marking Lilly's 150th anniversary, Ricks tied the new computing power to the company's broader research agenda, saying the supercomputer is already being used to help scientists solve complex problems and in manufacturing ^[2]. He also discussed Lilly's push into more powerful GLP-1 medicines meant to help people lose weight "faster and more safely," and specifically mentioned the company's oral GLP-1 drug, orforglipron, sold under the brand name Foundayo ^[2]. Ricks separately described work with Purdue University on automation and robotics for manufacturing, and said Lilly is expanding research into neurodegenerative diseases such as ALS, frontotemporal dementia and Lewy body dementia, an effort he said reflects lessons from the period after Lilly lost patent protection on Prozac ^[2].

Why it matters for patients

This is a corporate infrastructure and research announcement, not news about a new drug or a change to an existing prescription. Lilly says the computing power is meant to speed up the search for new medicines and shorten development timelines, but the company's own press release cautions that there is no guarantee the supercomputer will produce commercially successful products ^[1]. For patients currently using or considering GLP-1 medicines like Zepbound, Mounjaro, or Foundayo, nothing about dosing, availability, side effects, or cost changes because of this announcement.

The more concrete link to GLP-1 drugs is indirect: Ricks connected the computing investment to Lilly's stated goal of developing more powerful weight-loss and diabetes medicines and highlighted the recently launched oral drug orforglipron ^[2]. Whether or how the new supercomputer speeds up any specific GLP-1 research project is not detailed in the sources.

What happens next

Lilly's presentation on the supercomputer, titled "Enterprise-Scale AI for Drug Discovery: Strategy, Infrastructure and Outcomes," was scheduled for October 28, 2025, at NVIDIA's AI conference in Washington, D.C. ^[1]. No further timeline for specific drug candidates or research outcomes tied to the system is given in the available sources, and it is not yet known when or whether the new infrastructure will lead to identifiable changes in Lilly's drug pipeline.