Relation Therapeutics expanded its AI-driven drug discovery relationship with GSK under a new deal potentially worth up to $110 million. The collaboration centers on generating large-scale cellular perturbation datasets for training cellular foundation models, using automated labs to produce high-throughput multi-omic readouts. Relation also unveiled MORGAN, a general-purpose cellular foundation model intended to predict how cells respond to genetic and pharmacological interventions across cell types. MORGAN’s tissue-specific models will be used as part of the collaboration, with Relation expected to generate petascale datasets through its integrated automation. The announcement reinforces that AI in drug discovery increasingly depends on proprietary, consistent perturbation data assets—not only model algorithms—and that pharma is willing to pay for dataset generation aligned to modeling objectives.