Deepcell and Gilead Sciences entered a codevelopment partnership to build a custom AI foundation model for Chinese hamster ovary (CHO) cell line development using single-cell morphology. The collaboration will use Deepcell’s REM-I platform to image Gilead’s CHO cell lines and train a self-supervised model designed to characterize morphological structure without manual labeling. The companies said the model aims to improve how teams select cell lines with desirable production attributes such as titer, stability, aggregation, and metabolic profiles. Deepcell also stated it expects to evaluate the first model on held-out Gilead cell lines and plans to publish an open-weight version for broader access. For the bioprocessing workflow, the partnership targets a persistent bottleneck in cell line development—turning morphology information into early, scalable decisions that typically depend on labor-intensive manual review and experimental iteration.