NVIDIA's BioNeMo platform is becoming the de facto AI backbone for pharmaceutical R&D. Eli Lilly and Thermo Fisher have both signed co-innovation lab agreements anchored to the platform.1
The move signals a structural shift: major pharma companies are no longer piloting AI — they are building production infrastructure around it.
Four specialized AI biotech startups — Basecamp Research, Boltz Lab, Owkin, and Edison Scientific — simultaneously released new foundation models.1 The coordinated release suggests the sector has reached an inflection point, moving from bespoke experiments to deployable, scalable tools.
Novo Nordisk's strategy offers the clearest illustration of how incumbents are repositioning. The company licensed its Parkinson's cell therapy program to AI-native partner Cellular Intelligence and closed its internal development unit.1 Novo Nordisk retains licensing upside while transferring early-stage biology risk to a partner better equipped to run it.
Novo Nordisk's stock rose 24.9% over the past month.1 Investors appear to be rewarding the strategic logic: shed cost-heavy internal programs, capture royalty streams, and redeploy capital toward the GLP-1 franchise where the company holds dominant market position.
The FDA granted Fast Track designation to the Parkinson's cell therapy program.1 That designation reduces one of the largest adoption barriers for AI-developed therapeutics — regulatory uncertainty — and reinforces investor confidence across the broader AI-pharma cohort.
The emerging model is a three-layer stack. Platform providers like NVIDIA supply the compute and model infrastructure. AI biotech startups supply specialized foundation models and biology expertise. Large pharma supplies capital, regulatory relationships, and commercial scale — while licensing out the earliest and most uncertain R&D stages.
For financial analysts tracking pharma R&D spend, the implication is a reallocation, not a reduction. Internal biology headcount may shrink. But partnership and licensing budgets will grow as incumbents compete to secure the most capable AI-native partners before they are absorbed or capacity-constrained.
The companies that move fastest to anchor relationships with leading AI biotech platforms — or co-develop proprietary infrastructure as Lilly and Thermo Fisher are doing with NVIDIA — will hold a durable R&D cost and speed advantage over those that remain dependent on legacy discovery workflows.
