Microbe-mediated arsenic redox transformation and methylation are long-recognized to control its molecular-level speciation, thus bioavailability to rice plants. Arsenic thiolation, i.e. the formation of soluble inorganic thioarsenates (AsVS-IInO4-n3-, with n=1-4) and methylated thioarsenates (AsVS-IIn(CH3)O3-n3- or AsVS-IIn(CH3)2O2-n3- with n=1-2), has to date not been considered for As cycling in paddy fields mainly due to analytical limitations. Here we prove for the first time the widespread and quantitative importance of arsenic thiolation in paddy soils in Italy, France and China, combining evidence from field-based porewater sampling, lab soil incubations and mesocosm rice cultivation. The share of thiolated As ranged from 9% in non-amended field sites to 34% in sulfate-spiked incubation experiments and negatively correlated with soil pH on a large geographical scale. The reason for this negative correlation was the dominance of methylated thioarsenates (up to 30% of total arsenic) which dominate in acidic paddy soils, while inorganic thioarsenates (up to 18% of total arsenic) dominate in neutral and alkaline paddy soils. Furthermore, we prove the survival and significance of thioarsenates during the whole rice growing season, with enhancement of its contribution by sulfate-containing fertilizers, especially in soils low in organic carbon. Given the ubiquitous natural occurrence of arsenic and sulfur, we suggest that arsenic thiolation needs to be considered in global arsenic budgets not only in paddy soils, but also in other analogous sulfidogenic habitats (e.g., floodplains, peatlands, wetlands). Arsenic thiolation, its geographic distribution as well as its dependence on biogeochemical factors need to be deciphered in order to paint a more accurate picture of global arsenic cycling.