Novo Nordisk's stock rose 24.9% after the Danish drugmaker simultaneously closed its internal cell therapy unit and licensed its Parkinson's disease program to Cellular Intelligence, an AI-enabled biotech.1
The market's reaction signals a clear preference: asset-light, AI-accelerated models over in-house wet-lab operations. Novo Nordisk's pivot is among the most visible in European pharma so far.
NVIDIA's BioNeMo platform is becoming the central infrastructure layer for this shift. It now connects established pharma names — including Novo Nordisk, Eli Lilly, and Thermo Fisher — to a growing ecosystem of AI-native drug discovery firms.1
For European companies, the competitive pressure is structural, not cyclical. AI is no longer an experimental add-on; it is becoming foundational infrastructure for target identification, molecule design, and clinical trial optimisation.
The pace of model development is accelerating. Several specialised biological foundation models have launched in quick succession: Basecamp Research's EDEN, Owkin's OwkinZero, Boltz Lab's platform, Edison Scientific's Kosmos, and Natera's genomics model.1 Each targets a distinct slice of drug discovery, pushing competitive differentiation up the stack from generic AI capability to domain-specific performance.
Thermo Fisher, with its broad European laboratory services footprint, is positioned as both a supplier and an adopter within this infrastructure buildout. Pharma clients are increasingly demanding AI-integrated workflows rather than standalone tools.
Novo Nordisk's move also reflects a broader reconfiguration of risk in drug development. Licensing late-stage assets to AI-capable partners transfers execution risk while retaining commercial upside — a model that investors are now rewarding explicitly.
Danish biotech is adapting. The closure of Novo Nordisk's cell therapy unit cuts costs tied to expensive wet-lab infrastructure, while the Cellular Intelligence deal preserves optionality in neurodegeneration — one of pharma's most difficult therapeutic areas.
The BioNeMo ecosystem's expansion suggests the platform layer is close to commoditisation. When infrastructure stabilises, the firms with the strongest domain-specific models and proprietary biological data will hold the durable edge.
For European pharma, the window to build or acquire those capabilities is narrowing.
