1. Soil microbes are fed primarily by root-derived substrates, fulfil functions such as mineralization, immobilization, decomposition, pathogeneity and improvement of plant nutrition, and form the basis of the below-ground food web. Hitherto, belowground processes have generally been monitored using a 'black-box' approach, thereby ignoring effects of global change at a finer level of resolution. We describe shifts in the activity between microbial functional groups associated with roots of Artemisia tridentata, and the influence of this change on higher trophic levels, 2. We tested the hypothesis that elevated atmospheric CO2 causes the soil community to change qualitatively. We measured the responses of several soil microbe and soil microfaunal parameters to a double-ambient CO2 concentration and nutrient additions. The soil community, as measured by those parameters, showed great changes in response to the treatments. There was a very strong interaction between elevated CO2 and the nutrient addition. 3. Under low nutrient conditions, total microbial biomass did not change under elevated atmospheric CO2, but doubled under conditions of elevated CO2 and added nutrients. As we increased the resolution of our analysis, however, results shifted. Under low nutrient conditions, mycorrhizal fungi responded positively to elevated CO2, whereas with added soil nutrients they responded negatively to the same elevated CO2 concentration. Bacteria and non-mycorrhizal fungi did not respond under the former conditions but more than doubled in biomass under conditions of elevated CO2 and added nutrients. Soil fauna was also affected by the treatments. Overall, elevated CO2 shifted carbon flow in the plant-soil system to a more mutualistic-closed, mycorrhizal-dominated system, whereas the combination of elevated CO2 and nutrient addition shifted carbon flow to a more opportunistic-open, saprobe/pathogen-dominated one. 4. This indicates that elevated atmospheric CO2 may lead to far less predictable feedback patterns than previously thought and that qualitative shifts in the soil community may be far more important than mere changes in total C sink strength.