Grazing intensity shifts plant–microbe competition for ammonium and nitrate in temperate grasslands

Abstract In temperate grasslands, competition between plants and soil microbes for inorganic nitrogen (N) is a critical process in regulating N retention, loss and overall ecosystem productivity. However, how grazing intensity alters this competition, particularly for specific N forms such as ammonium (NH 4 + ) and nitrate (NO 3 − ) remains poorly understood. We conducted a dual 15 N labelling experiment in a semi‐arid steppe in Inner Mongolia, China, where four grazing intensities (none, light, moderate and heavy) have been maintained for 9 years. Grazing increased soil pH, reduced soil moisture and organic carbon, and decreased plant biomass (particularly under heavy grazing). Moderate grazing shifted microbial biomass and composition, declining Proteobacteria and increasing Ascomycota. These changes altered plant and microbial N uptake. Plant N preference shifted from NH 4 + to NO 3 − with increasing grazing intensity, while microbial NO 3 − uptake peaked under moderate grazing but declined under heavy grazing. In contrast, microbial NH 4 + uptake increased nonlinearly according to a quadratic relationship, resulting in microbial dominance under heavy grazing. The microbial‐to‐plant uptake ratio for NH 4 + revealed a tipping point, beyond which microbial uptake dominated. Our results demonstrate that grazing intensity regulates plant–microbial N competition by driving ammonium towards microbial immobilization and nitrate towards plant uptake, mediated by changes in microbial community composition. Fungal immobilization under heavy grazing drove microbial NH 4 + dominance, creating a tipping point. Synthesis . These findings highlight a threshold response in below‐ground nutrient partitioning and suggest that moderate grazing sustains balanced plant–microbe interactions in grassland ecosystems. This study provides a mechanistic framework based on a fungal‐driven NH 4 + tipping point, revealing critical competition thresholds that are not detectable in bulk N measurements.