The year opened hot, with the first weeks of January packed with deal flow, mega-rounds, platform launches, and AI model deployments as JPM week got underway. Companies doubled down on AI partnerships and infrastructure: for example, Eli Lilly and NVIDIA announced a $1 billion, five-year joint AI lab in San Francisco, aimed at making computational models core drug R&D infrastructure.
Efforts to amass massive “virtual biology” datasets also made news with Tahoe Therapeutics, Arc Institute, and Chan Zuckerberg Biohub combined roughly 120M single‑cell profiles and 225k perturbations into what they position as the largest open resource for virtual cell modeling. Illumina unveiled a Billion Cell Atlas, built with AstraZeneca, Merck, and Lilly, to train models on standardized cell-line experiments. Observers, though, kept a critical eye on translation, noting that cell lines are not patients and these curated perturbation signals may not cleanly map to real clinical biology. A smaller but distinctive example of the “data and population specificity” theme was BioMed X and the Barbados government’s project to model early diabetic kidney disease using deep molecular profiling and AI, framed as a population-specific “digital African twin.”
Pharma interest in virtual-cell-style modeling showed up in two directions: GSK signed a five-year, non-exclusive license and collaboration with Noetik, committing $50M to use its virtual cell foundation models and bespoke spatial oncology datasets in non-small cell lung cancer and colorectal cancer, while Bristol Myers Squibb entered a strategic clinical-development agreement with Immunai to apply its immune “operating system” to high-dimensional trial data to decode patient immune responses/variability in treatment outcomes.
Another clear theme was the continued movement of AI from proofs-of-concept into everyday workflows. Pharmaceutical players started bringing AI in-house and embedding it into R&D operations – AstraZeneca’s mid-month acquisition of Modella AI was explicitly to fold foundation models and “agentic” AI workflows throughout its oncology pipeline. On the tools side, MIT spinout Boltz (the founder trio is behind the well-known BoltzGen model) launched as a public benefit corporation with $28M seed round, and a multi-year collaboration with Pfizer.
Below is a selective rundown of the early-year signals we tracked most closely.
