Deepcell and Gilead agreed to co-develop an AI foundation model for Chinese hamster ovary (CHO) cell line development using single-cell morphology data. Under the collaboration, Deepcell will image Gilead’s CHO lines on its REM-I platform and train a self-supervised model that can characterize morphological structure without manual labeling. Deepcell said the model is intended to help cell line development teams identify higher-quality lines earlier—supporting attributes such as titer, stability, aggregation, and metabolic profile. The companies expect to evaluate the first model on held-out Gilead cell lines and plan to publish an open-weight version for the broader bioprocessing community. The partnership aligns with a broader push to industrialize AI workflows in biomanufacturing, where morphology-based readouts can reduce time spent on iterative lab screening and improve consistency across CHO platforms.