Nvidia, now the world’s first $5 trillion company, is reshaping healthcare and research through a network of recent AI partnerships across pharma, life sciences and medtech.
The Lilly deal
The largest of them is with Eli Lilly. The two will train biomedical foundation and frontier models to speed up discovery, imaging, trials and even biomanufacturing. Nvidia’s infrastructure will be implemented through Lilly TuneLab, built on more than $1B of Lilly data and located in Lilly’s Indianapolis data center.
The goal is to train proprietary AI models for drug discovery, automated manufacturing and clinical-trial optimization, significantly reducing development and approval timelines that traditionally take a decade or more.
The rest of the network
Verily, part of Alphabet, is integrating Nvidia AI into its precision-health platform for personalized medicine. Johnson & Johnson is using Nvidia models to train surgical robots and improve real-time decision-making. Innovaccer is partnering with Nvidia to build multimodal AI for patient data, increasing insights across imaging, EHR and genomic data, and integrating AI at provider frontlines through ambient scribing, OCR and agentic workflows.
Nvidia is also backing academic initiatives: CodonFM for RNA medicine development, La-Proteina for 3-D protein-structure generation, medical imaging models for radiology and pathology AI, and a Chan Zuckerberg Initiative collaboration on virtual cell models for researchers worldwide.
The scale of the opportunity shows up in two numbers. On average, target discovery takes about five years traditionally and roughly one year with AI; preclinical work drops from about 3.5 years to 1.5; and clinical phases I through III fall from about 6.5 years to 3.5, according to data from McKinsey, Nature Biotechnology and Deloitte Life Sciences. Global investment in AI-driven drug discovery has risen from $0.9B in 2019 to $2.3B in 2021, $4.8B in 2023 and $9.5B in 2025, per NVIDIA, EvaluatePharma, CB Insights and PitchBook Biotech.
What happens next
Nvidia is not just entering healthcare. It is establishing itself as the backbone of biology’s next industrial revolution. Expect novel drugs to arrive more frequently, with each iteration cutting discovery costs and bringing drugs to market faster and potentially cheaper.
In the short term, through 2025 and 2026, pharma companies race to integrate Nvidia AI stacks, with early results in molecule prediction and manufacturing efficiency. In the mid term, from 2027 to 2030, AI factories become standard across more top pharmas globally and the drug-discovery cost curve drops 20-40%. Longer term, past 2030, the entire healthcare economy shifts toward predictive value chains, where AI designs the drug, tests it virtually and guides some surgical operations.
That raises questions worth sitting with. If computing accelerates biology this much, will the FDA adapt with a new AI fast-track review pathway? Who will own the IP, the company that designed the molecule or the AI that generated it? And if AI factories shorten development and approval timelines by 60-70%, will drug prices actually drop?
Keep a close eye on next year’s updates on AI integration and drug discoveries.