The cycling of material from Earth's surface environment into its interior can couple mantle oxidation state to the evolution of the oceans and atmosphere. A major uncertainty in this exchange is whether altered oceanic crust entering subduction zones can carry the oxidised signal it inherits during alteration at the ridge into the deep mantle for long-term storage. Recycled oceanic crust may be entrained into mantle upwellings and melt under ocean islands, creating the potential for basalt chemistry to constrain solid Earth–hydrosphere redox coupling. Numerous independent observations suggest that Iceland contains a significant recycled oceanic crustal component, making it an ideal locality to investigate links between redox proxies and geochemical indices of enrichment. We have interrogated the elemental, isotope and redox geochemistry of basalts from the Reykjanes Ridge, which forms a 700 km transect of the Iceland plume. Over this distance, geophysical and geochemical tracers of plume influence vary dramatically, with the basalts recording both long- and short-wavelength heterogeneity in the Iceland plume. We present new high-precision Fe-XANES measurements of Fe3+/∑FeFe3+/∑Fe on a suite of 64 basalt glasses from the Reykjanes Ridge. These basalts exhibit positive correlations between Fe3+/∑FeFe3+/∑Fe and trace element and isotopic signals of enrichment, and become progressively oxidised towards Iceland: fractionation-corrected Fe3+/∑FeFe3+/∑Fe increases by ∼0.015 and ΔQFM by ∼0.2 log units. We rule out a role for sulfur degassing in creating this trend, and by considering various redox melting processes and metasomatic source enrichment mechanisms, conclude that an intrinsically oxidised component within the Icelandic mantle is required. Given the previous evidence for entrained oceanic crustal material within the Iceland plume, we consider this the most plausible carrier of the oxidised signal. To determine the ferric iron content of the recycled component ([Fe2O3]source[Fe2O3]source) we project observed liquid compositions to an estimate of Fe2O3 in the pure enriched endmember melt, and then apply simple fractional melting models, considering lherzolitic and pyroxenitic source mineralogies, to estimate [Fe2O3](source)[Fe2O3](source) content. Propagating uncertainty through these steps, we obtain a range of [Fe2O3](source)[Fe2O3](source) for the enriched melts (0.9–1.4 wt%) that is significantly greater than the ferric iron content of typical upper mantle lherzolites. This range of ferric iron contents is consistent with a hybridised lherzolite–basalt (pyroxenite) mantle component. The oxidised signal in enriched Icelandic basalts is therefore potential evidence for seafloor–hydrosphere interaction having oxidised ancient mid-ocean ridge crust, generating a return flux of oxygen into the deep mantle.