Deepcell and Gilead Sciences announced a partnership to co-develop a custom AI foundation model for Chinese hamster ovary (CHO) cell line development using single-cell morphology data. The collaboration will leverage Deepcell’s REM-I platform and train using self-supervised learning to characterize morphological structure without manual labeling. Under the agreement, Deepcell will image Gilead’s CHO cell lines on REM-I and train a model designed to give earlier, more efficient readouts on cell-line quality attributes such as titer, stability, aggregation, and metabolic profiles. The companies said they expect to evaluate the initial model on held-out Gilead cell lines. Deepcell said the partners plan to publish an open-weight version of the model to make it broadly available to the bioprocessing community. Deepcell previously raised $73 million in Series B funding in early 2022, and it has commercialized label-free single-cell morphology analysis and sorting. For biopharma developers, the deal is a direct signal that AI foundation models are moving from discovery workflows into manufacturing-adjacent cell line development—where faster quality assessment can translate into timeline and cost improvements.