Lower-Friction AI Adoption in Pharma

This points to the core buying advantage of software style AI in drug discovery, it lets a pharma team add AI on top of tools and workflows they already trust, instead of entering a co development relationship that claims part of the molecule economics. Schrödinger’s 2026 Bunsen deal with Bristol Myers Squibb fits that pattern, while Isomorphic’s Novartis work sits closer to a collaboration model built around shared research programs.

• Schrödinger is easier for procurement to slot into an existing discovery stack because Bunsen is sold with a software agreement and runs alongside its established modeling tools. That means scientists can use AI inside familiar screens and validation loops, instead of standing up a new joint venture around targets and downstream milestones.
  1 schrodinger 5 sec
• Isomorphic’s partnerships are deeper and potentially more valuable per win, but they ask more from the customer. Its Novartis relationship began as a strategic research collaboration in January 2024 and expanded in February 2025, which implies tighter scientific integration and more negotiation over program scope, data sharing, and rights.
  3 isomorphiclabs 4 isomorphiclabs
• Biologics specialists show the other end of the market. Generate has pushed GB-0895 into Phase 3, proving an integrated build your own pipeline model can reach the clinic. Chai has taken the lighter software style route, signing multiple antibody design collaborations with large pharma, including Novartis and argenx, without taking on full clinical development load.
  6 generatebiomedicines 7 generatebiomedicines 8 chaidiscovery

The market is moving toward a split. Incumbent software vendors will win budget faster by fitting into existing lab workflows, while collaboration heavy AI biotechs will chase fewer, larger deals tied to actual drug creation. Over time, pharma will keep multi homing across both models, then concentrate spend on the platforms that turn model output into clinical candidates.