Given the high genetic heterogeneity of inherited retinal degenerations (IRDs), a wide applicable treatment would be desirable to halt/slow progressive photoreceptor (PR) cell loss in a mutation-independent manner. In addition to its erythropoietic activity, erythropoietin (EPO) presents neurotrophic characteristics. We have previously shown that adeno-associated viral (AAV) vector-mediated systemic EPO delivery protects from PR degeneration. However, this is associated with an undesired hematocrit increase that could contribute to PR protection. Non-erythropoietic EPO derivatives (EPO-D) are available which allow us to dissect erythropoiesis's role in PR preservation and may be more versatile and safe than EPO as anti-apoptotic agents. We delivered in animal models of light-induced or genetic retinal degeneration either intramuscularly or subretinally AAV vectors encoding EPO or one of the three selected EPO-D: the mutant S100E, the helix A- and B-derived EPO-mimetic peptides. We observed that (i) systemic expression of S100E induces a significantly lower hematocrit increase than EPO and provides similar protection from PR degeneration, and (ii) intraocular expression of EPO-D protects PR from degeneration in the absence of significant hematocrit increase. On the basis of this, we conclude that erythropoiesis is not required for EPO-mediated PR protection. However, the lower efficacy observed when EPO or S100E is expressed intraocularly rather than systemically suggests that hormone systemic effects contribute to PR protection. Unlike S100E, EPO-mimetic peptides preserve PR only when given locally, suggesting that different EPO-D have a different potency or mode of action. In conclusion, our data show that subretinal delivery of AAV vectors encoding EPO-D protects from light-induced and genetic PR degeneration.