Relation Therapeutics struck a collaboration with GSK to generate large-scale cellular perturbation datasets for use in cellular foundation models. Relation may receive up to $110 million in upfront and milestone payments, and the deal includes deployment of Relation’s MORGAN general-purpose cellular foundation model for predicting responses across genetic and pharmacological interventions. The agreement calls for petascale, time-resolved multi-omic perturbation data generated via automated laboratories, with the output intended to strengthen confidence in target discovery and validation. Relation said datasets will be built across relevant cell types rather than tuned to a single indication. The partnership adds to the growing “data infrastructure” playbook in AI drug discovery, where model performance is increasingly tied to reproducible, high-throughput experimental generation.