The rapidly rising concentration of atmospheric CO2 has the potential to alter forest and global carbon cycles by altering important processes that occur in soil. Forest soils contain the largest and longest lived carbon pools in terrestrial ecosystems and are therefore extremely important to the land–atmosphere exchange of carbon and future climate. Soil respiration is a sensitive integrator of many soil processes that control carbon storage in soil, and is therefore a good metric of changes to soil carbon cycling. Here, we summarize soil respiration data from four forest free-air carbon dioxide enrichment (FACE) experiments in developing and established forests that have been exposed to elevated atmospheric [CO2] (168 μL L−1 average enrichment) for 2–6 years. The sites have similar experimental design and use similar methodology (closed-path infrared gas analyzers) to measure soil respiration, but differ in species composition of the respective forest communities. We found that elevated atmospheric [CO2] stimulated soil respiration at all sites, and this response persisted for up to 6 years. Young developing stands experienced greater stimulation than did more established stands, increasing 39% and 16%, respectively, averaged over all years and communities. Further, at sites that had more than one community, we found that species composition of the dominant trees was a major controller of the absolute soil CO2 efflux and the degree of stimulation from CO2 enrichment. Interestingly, we found that the temperature sensitivity of bulk soil respiration appeared to be unaffected by elevated atmospheric CO2. These findings suggest that stage of stand development and species composition should be explicitly accounted for when extrapolating results from elevated CO2 experiments or modeling forest and global carbon cycles.