Cell competition in the thymus promotes turnover and functions as a tumor suppressor by inhibiting leukemia. Using thymus transplantation experiments, we have shown that the presence of T lymphocyte precursors, recently seeding the thymus, promotes the clearance of precursors with a longer time of thymus residency. If cell competition is impaired and no cells seed the thymus, the organ is capable of sustaining T lymphocyte production, a state termed thymus autonomy. However, we observed consistently that prolonged autonomy is permissive to the emergence of T cell acute lymphoblastic leukemia (T-ALL). This resembled the onset of T-ALL in patients treated by gene therapy for X-linked severe combined immunodeficiency (SCID-X1). Following treatment, thymus activity was established, with T lymphocyte production, although no bone marrow contribution was detected. However, some patients developed T-ALL. The favored explanation for malignant transformation was considered to be genotoxicity due to integration of the retroviral vector next to oncogenes, thereby activating them ectopically. Although plausible, we consider an alternative, mutually nonexclusive explanation: that any condition enabling prolonged thymus autonomy will promote leukemogenesis. In support of this view, two independent studies have recently shown that the efficacy of reconstitution of the bone marrow in the context of SCID-X1 dramatically influences the outcome of treatment, and that lymphoid malignancies emerge following transplantation of a small number of healthy progenitors. Here, we discuss the most recent data in light of our own studies in thymopoiesis and the conditions that trigger malignant transformation of thymocytes in various experimental and clinical settings.