# $1B over five years. NVIDIA + Eli Lilly and Company.

> NVIDIA and Eli Lilly are putting $1B into a co-located Bay Area lab where Lilly scientists and NVIDIA engineers build drug discovery models side by side. Not a typical AI partnership. The real story is closed-loop discovery, and the moat is the loop, not the model.

URL: https://www.ch-healthtech.com/insights/1b-over-five-years-nvidia-eli-lilly-company
Markdown: https://www.ch-healthtech.com/insights/1b-over-five-years-nvidia-eli-lilly-company.md
Published: 2026-01-22
Updated: 2026-05-06
Author: Christian Hein
Tags: technology/artificial-intelligence, technology/machine-learning, function/innovation-management, industry/large-pharma, activity/partnership-alliance, function/strategic-planning, function/pre-clinical-research

---


## TL;DR

$1B over five years. NVIDIA and Eli Lilly are building a co-located Bay Area lab where Lilly scientists and NVIDIA engineers build drug discovery models side by side. Closed-loop discovery: wet lab runs, data lands, models retrain, next experiment gets designed, repeat 24/7. AI is becoming basic R&D infrastructure, the way supercomputing used to be. The moat isn’t the model. It’s the loop. If you’re Pfizer, Roche, or Sanofi: do you build, buy, or rent this capability? Because partnering with an AI vendor won’t close the loop.

$1B over five years. NVIDIA + Eli Lilly and Company.

Not a typical “AI partnership”. A co-located lab in the Bay Area where Lilly scientists and NVIDIA engineers build drug discovery models side by side.

The real story: closed loop discovery. Wet lab runs. Data lands. Models retrain. Next experiment gets designed. Repeat, 24/7.

Built on NVIDIA’s next-gen Vera Rubin chipset, plus robotics and “physical AI” to scale the lab work.

NVIDIA has been everywhere in biotech (Recursion, Novo Nordisk, Illumina, Mayo). But this feels like a line in the sand: AI is becoming basic R&D infrastructure, the way supercomputing used to be.

Called this one in my 2026 predictions. The moat isn’t the model. It’s the loop.

If you’re Pfizer, Roche, or Sanofi: Do you build this capability, buy it, or rent it?

Because ‘partnering with an AI vendor’ won’t close the loop.

## Key takeaways

- $1B over five years between NVIDIA and Lilly. A co-located Bay Area lab where scientists and engineers build models together, not an arm’s-length partnership.
- The architecture is what matters: closed-loop discovery. Wet lab runs feed data back into models, which design the next experiment, continuously.
- Built on NVIDIA’s next-gen Vera Rubin chipset plus robotics and “physical AI” to scale the lab work itself.
- AI is becoming basic R&D infrastructure, the same way supercomputing did. NVIDIA has been laying the groundwork everywhere (Recursion, Novo Nordisk, Illumina, Mayo).
- The competitive moat in drug discovery is not which foundation model you use. It is whether you have built the loop.
- For Pfizer, Roche, Sanofi and others watching from the sidelines: build, buy, or rent. Partnering with an AI vendor won’t close the loop.