Mineral‐associated soil organic matter (MAOM) is the largest, slowest cycling pool of carbon (C) in the terrestrial biosphere. MAOM is primarily derived from plant and microbial sources, yet the relative contributions of these two sources to MAOM remain unresolved. Resolving this issue is essential for managing and modeling soil carbon responses to environmental change. Microbial biomarkers, particularly amino sugars, are the primary method used to estimate microbial versus plant contributions to MAOM, despite systematic biases associated with these estimates. There is a clear need for independent lines of evidence to help determine the relative importance of plant versus microbial contributions to MAOM. Here, we synthesized 288 datasets of C/N ratios for MAOM, particulate organic matter (POM), and microbial biomass across the soils of forests, grasslands, and croplands. Microbial biomass is the source of microbial residues that form MAOM, whereas the POM pool is the direct precursor of plant residues that form MAOM. We then used a stoichiometric approach—based on two‐pool, isotope‐mixing models—to estimate the proportional contribution of plant residue (POM) versus microbial sources to the MAOM pool. Depending on the assumptions underlying our approach, microbial inputs accounted for between 34% and 47% of the MAOM pool, whereas plant residues contributed 53%–66%. Our results therefore challenge the existing hypothesis that microbial contributions are the dominant constituents of MAOM. We conclude that biogeochemical theory and models should account for multiple pathways of MAOM formation, and that multiple independent lines of evidence are required to resolve where and when plant versus microbial contributions are dominant in MAOM formation. The stoichiometry provides a new independent source of data for evaluating the relative contributions of plant and microbial inputs to mineral‐associated soil organic matter formation (MAOM). In our global databases, MAOM C/N ratios are greater than microbial C/N ratios and lower than POM C/N ratios across all ecosystem types. This suggest that MAOM contains both microbial and plant residues. Using fractional abundance of N N/(C + N) in POM and microbial biomass as end‐members in the two‐pool mixing model, we found that microbial inputs accounted for 34%–47% of the MAOM pool, whereas plant residues contributed 53%–66%. Our results challenge the increasingly popular view that microbial contributions are the dominant component of MAOM.