Soil pH is a key predictor of plant species occurrence owing to its effect on the availability of nutrients and phytotoxic metals. Although regional differences in realized soil pH niche ('niche shifts') have been reported since the 19th century, no study has disentangled how they are influenced by spatial differences in substrate availability, macroclimate, and competitors. We linked plot-level data on species occurrence and measured soil pH from dry grasslands in eight regions across Eurasia (n = 999 plots), spanning a geographic gradient of 6862 km. We calculated regional shifts in niche optimum (D_opt) and width (D_width) for 73 Species × Region 1 × Region 2 combinations (SRRs; 38 study species) using extended Huisman-Olff-Fresco models. Next, we used commonality analysis to partition the contribution of substrate availability, precipitation, and species traits indicative of competitive ability to variation in regional niche shifts. Shifts in optimum were rare (5% of SRRs with D_opt ≥ 1 pH units) but many species did not show optima within regions. By contrast, shifts in niche width were common (22% of SRRs with D_width ≥1 pH units) and there were pronounced interspecific differences. Whereas none of the three predictors significantly explained shifts in niche optimum, common and unique effects of substrate availability and precipitation accounted for 85% of variation in niche width. Our results suggest that substrate availability and precipitation could be the driving factors behind species regional shifts in niche width. Studies that address additional factors, such as other edaphic niches, and their variability at the regional and micro-scale will improve our understanding of the mechanisms underlying species distributions.