Novo Nordisk is closing its internal cell therapy unit and licensing a Parkinson's disease asset to the AI-driven Cellular Intelligence platform.3 The Danish drugmaker's shares gained 24.9% over the past 30 days, alongside strong core earnings.3 The pairing signals investors read AI-driven R&D outsourcing as additive, not a retreat.
The shift comes as a cluster of 'bio-native AI' foundation-model platforms position themselves as the new infrastructure layer for drug discovery. NVIDIA's BioNeMo, Basecamp Research's EDEN, Boltz Lab, Owkin's OwkinZero and Edison Scientific's Kosmos are each building models trained directly on biological data rather than repurposed language models. NVIDIA is pushing to become the connective tissue linking these efforts, striking lab-in-the-loop partnerships with Thermo Fisher and Eli Lilly.
Nasdaq-listed MindWalk Holdings is staking its position on patent protection for its underlying representation layer. The company's HYFT Technology encodes roughly 660 million biological patterns, described as universal fingerprints capturing conserved relationships between sequence, structure and function.2 MindWalk argues that representing biology itself, not just modeling it, is the harder problem and the more durable source of advantage.2
MindWalk has now filed a second patent, EP26187897.9, which the company says adds a distinct computational layer on top of its original HYFT patent, EP3881326A1.1 "It is the computational layer built on the original foundation, not a re-filing of it," the company said.1 The filing organizes biological patterns into a form reusable across MindWalk's infrastructure, customer programs and AI workflows.1
For European pharma, the Novo Nordisk case is the clearest signal yet. A top-tier incumbent is cutting an in-house therapeutic modality while outsourcing a specific disease program to an external AI platform, and the market is rewarding it. As NVIDIA and specialist AI-biology firms build out shared infrastructure for drug discovery, European drugmakers face pressure to decide which R&D functions to keep internal and which to license out.
