Apoptosis has been established as a mechanism of anti-cancer defense. Members of the BCL-2 family are critical mediators of apoptotic cell death in health and disease, often found to be deregulated in cancer and believed to lead to the survival of malignant clones. However, over the years, a number of studies pointed out that a model in which cell death resistance unambiguously acts as a barrier against malignant disease might be too simple. This is based on paradoxical observations made in tumor patients as well as mouse models indicating that apoptosis can indeed drive tumor formation, at least under certain circumstances. One possible explanation for this phenomenon is that apoptosis can promote proliferation critically needed to compensate for cell loss, for example, upon therapy, and to restore tissue homeostasis. However, this, at the same time, can promote tumor development by allowing expansion of selected clones. Usually, tissue resident stem/progenitor cells are a major source for repopulation, some of them potentially carrying (age-, injury- or therapy-induced) genetic aberrations deleterious for the host. Thereby, apoptosis might drive genomic instability by facilitating the emergence of pathologic clones during phases of proliferation and subsequent replication stress-associated DNA damage. Tumorigenesis initiated by repeated cell attrition and repopulation, as confirmed in different genetic models, has parallels in human cancers, exemplified in therapy-induced secondary malignancies and myelodysplastic syndromes in patients with congenital bone marrow failure syndromes. Here, we aim to review evidence in support of the oncogenic role of stress-induced apoptosis.