The most telling moments at this year’s JPM Healthcare Conference were not the loud announcements—they were the structural signals embedded within them.
No blockbuster M&A headlines. No theatrical pipeline unveilings.
And yet, for R&D leaders paying attention, the message was unambiguous: pharma R&D is entering an infrastructure reset.
AI is no longer accelerating R&D—it is redefining it
The collaboration between Eli Lilly and Nvidia, a new AI co-innovation lab with up to $1 billion committed over five years — is significant not primarily because of its dollar figure. Still, because of its structure and strategic design, it co-locates AI engineers and pharmaceutical experts, builds a continuous scientist-in-the-loop drug discovery workflow, and tightly integrates AI model training with laboratory experimentation, creating a new blueprint that could reshape how medicines are discovered and developed.
This is not a software deployment.
It is the industrialization of AI as core R&D infrastructure—on par with automation, wet-lab robotics, or high-throughput screening.
The implication is profound:
Competitive advantage will no longer come from access to AI, but from how deeply AI is embedded into scientific decision-making—from target selection and molecular design to trial simulation and attrition forecasting.
For R&D leaders, this marks a shift from “AI adoption” to AI-native operating models.
Speed is being measured differently
The acquisition of Boston-based Modella AI by AstraZeneca reflects a strategic evolution in how the company defines R&D speed — shifting from external experimentation with AI tools toward fully integrating advanced multimodal AI models and talent into its oncology research operations to accelerate clinical development, biomarker discovery, and data-driven decision-making across its pipeline.
This is not about shaving months off development timelines downstream.
It is about compressing uncertainty upstream.
AI-driven discovery is being used to:
- Improve early confidence in biological hypotheses
- Reduce late-stage failure by improving early signal quality
- Shift R&D economics toward earlier, smarter decisions
In practice, boards are not asking R&D to move faster—they are asking it to know sooner.
The absence of mega-M&A is a sign of scientific discipline, not hesitation
This JPM surprised many by what didn’t happen: no major M&A announcements.
But deals did happen—just differently.
In January 2026, Novartis entered into strategic licensing and collaboration agreements with SciNeuro Pharmaceuticals and Zonsen PepLib Biotech, granting it exclusive worldwide rights to advance SciNeuro’s Alzheimer’s antibody program and a peptide-based radioligand therapy, respectively. These deals involve upfront payments ($165 M to SciNeuro; $50 M to PepLib) plus potential milestone and royalty payments, reflecting targeted collaboration structures that combine early-stage partnership with milestone-linked economic incentives. signal a preference for:
- Modality-specific bets (peptides, neuro, AI-enabled platforms)
- Scientific de-risking before organizational integration
- Flexibility over scale
For R&D leaders, this means scientific rigor—not deal size—is now the primary currency in partnering discussions.
Policy is quietly reshaping R&D expectations
Off-the-record remarks by leaders such as Marty Makary and Mehmet Oz reinforced a reality that R&D teams can no longer treat as downstream noise.
Pricing pressure, value-based frameworks, and national health priorities are increasingly shaping what gets developed, not just how it gets approved.
This elevates the role of R&D in:
- Generating real-world evidence earlier
- Designing programs with clear health-economic narratives
- Supporting access and reimbursement discussions from the outset
R&D is being pulled closer to policy, not shielded from it.
Data is moving from observation to intervention
Partnerships like Oura with Fullscript highlight another shift: patient-generated data is crossing the line from monitoring into clinical relevance.
For R&D, this opens new possibilities—and responsibilities:
- Digital biomarkers influencing trial design
- Continuous data streams informing endpoints
- Greater complexity in data governance and validation
The opportunity is real, but so is the need for scientific and regulatory discipline.
The strategic takeaway for pharma R&D leaders
JPM 2026 made one thing clear:
R&D excellence is no longer defined solely by scientific depth, but by decision intelligence.
The most advanced organizations are:
- Embedding AI into the fabric of R&D, not layering it on top
- Prioritizing early certainty over late acceleration
- Treating partnerships as scientific options, not corporate events
- Aligning discovery, evidence, and value narratives from day one
This is not a reinvention of R&D—but it is a recalibration of what leadership in R&D now requires.
