NVIDIA's BioNeMo has become foundational AI infrastructure for global pharmaceutical R&D, anchoring partnerships with Eli Lilly and Thermo Fisher and powering specialized models at Terray Therapeutics and Apheris.1
BioNeMo delivers pre-trained biological models that companies fine-tune for specific drug targets. Early adopters are compressing timelines that once required years of wet-lab iteration across research centers worldwide.
Denmark's Novo Nordisk — Europe's most valuable company — closed its internal cell therapy unit and licensed Parkinson's disease assets to AI-native Cellular Intelligence.1 The decision reflects a calculation now spreading across global pharma: AI-specialized partners are faster and cheaper than internal capability.
Novo Nordisk is concentrating on GLP-1, its core metabolic franchise, while exiting disease areas where AI platforms hold comparative advantage. The externalization model is accelerating across the US, Europe, and Asia as platform biology matures.
Platform launches are clustering. Boltz, France's Owkin, Basecamp EDEN, Edison Kosmos, and Natera all launched or expanded AI drug discovery platforms recently.1 The concentration of launches marks a shift from experimental tools to production-grade infrastructure supporting clinical pipelines.
Investor sentiment is bullish. Conviction in leading names like Novo Nordisk remains strong, and the overall outlook for AI-driven drug discovery continues to improve.1
The architecture solidifying across the industry runs three layers: NVIDIA at the compute base, BioNeMo-powered foundation models in the middle, and application-layer biotech firms on top. Companies at every layer are locking in partnerships before the market consolidates around dominant platforms.
Drug discovery has historically failed at high rates over decade-long timelines. AI platforms attacking target identification, molecule generation, and clinical trial design simultaneously — rather than one bottleneck — change the calculus fundamentally.
The competitive question is which foundation model providers achieve the validated results needed to become default biology infrastructure — the equivalent of AWS or Azure for drug development. The race is global, and the window to establish platform dominance is narrowing.
