Much information is accumulating on the effect of cerium oxide nanoparticles (CNPs) as cell-protective agents, reducing oxidative stress through their unique ability of scavenging noxious reactive oxygen species via an energy-free, auto-regenerative redox cycle, where superoxides and peroxides are sequentially reduced exploiting the double valence (Ce3+/Ce4+) on nanoparticle surface. In vitro and in vivo studies consistently report that CNPs are responsible for attenuating and preventing almost any oxidative damage and pathology. Particularly, CNPs were found to exert strong anticancer activities, helping correcting the aberrant homeostasis of cancer microenvironment, normalizing stroma-epithelial communication, contrasting angiogenesis, and strengthening the immune response, leading to reduction of tumor mass in vivo. Since these homeostatic alterations are of an oxidative nature, their relief is generally attributed to CNPs redox activity. Other studies however reported that CNPs exert selective cytotoxic activity against cancer cells and sensitize cancer cells to chemotherapy- and radiotherapy-induced apoptosis: such effects are hardly the result of antioxidant activity, suggesting that CNPs exert such important anticancer effects through additional, non-redox mechanisms. Indeed, using Sm-doped CNPs devoid of redox activity, we could recently demonstrate that the radio-sensitizing effect of CNPs on human keratinocytes is independent from the redox switch. Mechanisms involving particle dissolution with release of toxic Ce4+ atoms, or differential inhibition of the catalase vs. SOD-mimetic activity with accumulation of H2O2 have been proposed, explaining such intriguing findings only partially. Much effort is urgently required to address the unconventional mechanisms of the non-redox bioactivity of CNPs, which may provide unexpected medicinal tools against cancer.