Soil organic carbon is chemically heterogeneous, and microbial decomposers face a physiological challenge in metabolizing the diverse array of compounds present in soil. Different classes of polymeric compounds may require specialized enzymatic pathways for degradation, each of which requires an investment of microbial resources. Here we tested the resource allocation hypothesis, which posits that decomposition rates should increase once substrate concentrations are sufficient to overcome biochemical investment costs. We also tested the alternative hypothesis that mixing different substrates increases resource acquisition through priming effects involving generalist enzymes. Using a microcosm approach, we varied the soil concentration of seven distinct substrates individually and in mixture. We found that the percent carbon respired from starch, cellulose, chitin, and the mixture was significantly reduced at the lowest substrate concentration. The activities of β-glucosidase and N-acetyl-glucosaminidase that target cellulose and chitin, respectively, were also significantly lower at the lowest concentrations of their target substrates. However, we did not observe parallel declines in enzyme activity with starch or the mixture. Some enzymes, such as β-xylosidase, were consistent with specialist strategies because they showed the highest activity in the presence of their target substrate. Other enzymes were more generalist, with activity observed across multiple substrates. Together, these results suggest that the costs of biochemical machinery limit microbial decomposition of substrates at low concentration. The presence of enzymes with low substrate specificity was not sufficient to overcome this constraint for some substrates. Concentration constraints driven by microbial allocation patterns may be common in mineral soil and could be represented in new biogeochemical models based on microbial physiology.