AbstractA positive soil carbon (C)‐climate feedback is embedded into the climatic models of the IPCC. However, recent global syntheses indicate that the temperature sensitivity of soil respiration (RS) in drylands, the largest biome on Earth, is actually lower in warmed than in control plots. Consequently, soil C losses with future warming are expected to be low compared with other biomes. Nevertheless, the empirical basis for these global extrapolations is still poor in drylands, due to the low number of field experiments testing the pathways behind the long‐term responses of soil respiration (RS) to warming. Importantly, global drylands are covered with biocrusts (communities formed by bryophytes, lichens, cyanobacteria, fungi, and bacteria), and thus,RSresponses to warming may be driven by both autotrophic and heterotrophic pathways. Here, we evaluated the effects of 8‐year experimental warming onRS, and the different pathways involved, in a biocrust‐dominated dryland in southern Spain. We also assessed the overall impacts on soil organic C (SOC) accumulation over time. Across the years and biocrust cover levels, warming reducedRSby 0.30 μmol CO2 m−2 s−1(95% CI = −0.24 to 0.84), although the negative warming effects were only significant after 3 years of elevated temperatures in areas with low initial biocrust cover. We found support for different pathways regulating the warming‐induced reduction inRSat areas with low (microbial thermal acclimation via reduced soil mass‐specific respiration and β‐glucosidase enzymatic activity) vs. high (microbial thermal acclimation jointly with a reduction in autotrophic respiration from decreased lichen cover) initial biocrust cover. Our 8‐year experimental study shows a reduction in soil respiration with warming and highlights that biocrusts should be explicitly included in modeling efforts aimed to quantify the soil C–climate feedback in drylands.