Illumina announced the launch of its Billion Cell Atlas, a genome‑wide perturbation dataset that the company says will reach one billion single‑cell measurements and support AI‑driven drug discovery. Illumina named AstraZeneca, Merck and Eli Lilly as founding participants and said the dataset will map cellular responses to genetic perturbations across disease‑relevant cell lines. The company framed the atlas as a training ground for large AI models and a validation resource for target biology. The initial release covers hundreds of millions of cells and perturbations; Illumina intends to scale to a billion cells over the next few years. Pharma partners said they will use the resource for target validation, training foundation models and building “virtual cell” systems. For industry readers: the dataset promises standardized, perturbation‑heavy training data that could shorten early discovery cycles and de‑risk target selection for small molecules, biologics and modality‑agnostic AI models. Technical note: the atlas couples systematic genetic perturbations with single‑cell readouts; these data are intended to capture transcriptional and phenotypic consequences of gene knockouts and activations that are directly usable to train machine‑learning models.