The soil component of atmospheric CO2 oxygen isotope budget is evaluated in light of the recent recognition of abiotic oxygen isotope exchange of atmospheric CO2 and soil water in excess of soil respiratory CO2 flux. By using variations of published analytical models, we estimate that the amount of atmospheric CO2 that undergoes oxygen isotope exchange with soil water, exclusive of soil-respired CO2, to be approximately 0.2--0.7 &mgr;mol m-2 s-1 for representative unsaturated soils from a range of biomes. Globally, the amount of atmospheric CO2 that undergoes oxygen isotope exchange with soil water through purely nonbiological processes is probably significantly larger than the current annual fossil fuel combustion, yet this process has been neglected in all recent global 18O-CO2 budgets. Furthermore, abiotic oxygen isotope exchange with soils will occur roughly equally in soils with low and high respiration rates, suggesting that soils with low respiration rates are currently under-represented disproportionately in the existing global CO2 oxygen isotope budgets. Because soils with low respiration rates tend to have the most extreme soil water &dgr;18O values, their underrepresentation may have a large and heretofore unsuspected impact on the global atmospheric C18O16O budget. Finally, soil carbon is no longer in steady state due to land use practices, and this additional source of CO2 to the atmosphere may contribute to the decreasing trend in atmospheric CO2--&dgr;18O values through time. ¿ 2001 American Geophysical Union