Eli Lilly and Nvidia, two primary drivers of the recent stock market rally, are deepening their alliance to connect silicon with biology through artificial-intelligence-powered drug development.
The two companies will establish a new AI co-innovation lab with a commitment to invest more than $1 billion over five years, Nvidia announced at the 2026 J.P. Morgan Healthcare Conference in San Francisco.
“This co-innovation lab is dedicated to achieving accelerated, closed-loop discovery and the creation of AI models to improve clinical development,” Kimberly Powell, vice president of healthcare at Nvidia, said in a press briefing.
Scientists from the two firms will come together at a new site in the Bay Area that the pair expects to open by the end of March, Powell said.
Both sides are allocating dedicated, incremental resources to the partnership. Lilly will contribute its drug R&D expertise and laboratory infrastructure, while Nvidia will provide AI capabilities including open biology models, multimodal foundation models, agentic and physical AI, and DGX Cloud capacity, according to Powell.
“We see this as a catalyst for the capabilities that will define the next era of drug discovery,” Diogo Rau, executive vice president and chief information and digital officer at Lilly, said in a Jan. 12 statement. “By working with NVIDIA, we’re uniting massive compute, specialized talent and the ability to shape data at immense scale. We’re moving toward a future where discovery is driven by rapid experimentation and increasingly customized models—an approach that reflects our commitment to leading applied AI in drug discovery and investing deeply in new forms of data generation and model development.”
The latest collaboration between Lilly and Nvidia builds on the companies’ “AI factory” project, under which Nvidia is helping Lilly build what could become the pharmaceutical industry’s largest supercomputer designed to accelerate drug discovery and shorten development timelines.
For the new joint lab, a major focus will be on producing real-world lab data to train and validate AI models.
“We’re going to continue to create ground truth data in the lab so that we can train these biology foundation models with multimodal data, taking advantage of all of the advancements that we’ve seen in [large language model] in biology,” Powell said.
With the help of AI, the team will work to shorten the process between hypotheses and discovery in the hopes of establishing a new scientific paradigm, Powell added.
In addition to R&D, the two firms will have the opportunity to apply AI across Lilly’s businesses, including manufacturing and commercial operations. Powell highlighted Nvidia’s three physical AI and robotics platforms—Omniverse, Isaac and Jetson—which she said can fit into Lilly’s manufacturing facilities.
Lilly and Nvidia are broadening their collaboration as both invest heavily in each other’s territories.
“Lilly is shifting from using AI as a tool to embracing it as a scientific collaborator,” Lilly’s chief AI officer, Thomas Fuchs, said in an October release about the new “AI factory.”
To extend the reach of its AI capabilities, Lilly last year launched Lilly TuneLab, an open innovation platform that gives biotech companies access to its AI drug discovery models. In return, the ecosystem creates a two-way flow of information that provides Lilly with new data to refine those models.
Friday, Schrödinger announced it’s integrating Lilly’s Tunelab into its drug design software, LiveDesign.
For its part, Nvidia will announce at JPM an expansion of its BioNeMo into a full open development platform for biology and drug discovery. BioNeMo includes several open models, which come with open datasets. Among some of the latest additions are a model for predicting 3D structure of RNA molecules, a reasoning model called ReaSyn v2 for ensuring AI-designed molecules are practical to synthesize and another for predicting toxicity, according to Powell. Besides, BioNeMo Recipes will help scale biological foundation model training, customization and deployment.
“Instead of just sending a model over the fence, we’re giving the whole developer community the end-to-end tools accompanied with the data sets, the models themselves, the training recipes and everything in between, so that the full AI life cycle can be taken advantage of and made very easy for the domain experts across the field,” Powell said.
On the robotics front, Multiply Labs, which offers robotic systems for manufacturing gene-modified cell therapies, is using Nvidia’s three physical AI platforms to create digital twins of entire labs, replacing manual, contamination-prone work with autonomous systems. According to Powell, the robotic systems could cut cell therapy manufacturing costs by 70% per dose and increase dose throughput 100-fold per square foot by reducing the need for highly trained personnel and related equipment.
At JPM, Nvidia also announced it is working with Thermo Fisher Scientific to build autonomous lab infrastructure.
“Laboratory automation is not just operational improvement,” Powell said. “It’s a foundational enabler of more intelligent biological AI.”