Retinoic acid receptor γ (RARγ) is selectively expressed by normal stem cells and is needed to maintain them. This review argues that cancer cells, particularly cancer stem cells, have hijacked this role. RARγ is overexpressed in cholangiocarcinoma and in colorectal, head and neck, hepatocellular, ovarian, pancreatic, prostate, and renal cancers, and in acute myeloid leukaemia it appears as a fusion protein. High expression is linked to high-grade disease, metastasis, and poor prognosis. Cancer cells often live in a low retinoic acid environment, where RARγ is the receptor activated first, and they depend on it for growth and survival. Antagonising RARγ, or all RARs, causes growth arrest and often cell death in leukaemia, hepatocellular, pancreatic, paediatric brain, and prostate cancer models. In prostate cancer, this includes necroptosis of cancer stem-like cells and patient-derived cells, with normal prostate cells less sensitive. Animal studies of a pan-RAR antagonist showed no major toxicity other than reversible inhibition of spermatogenesis, supporting RARγ as a promising target for eradicating cancer stem cells and addressing metastasis and relapse.