The standard approach for geologic storage of CO2 consists of injecting it as a supercritical CO2 phase. This approach places stringent requirements on the caprock, which must display: (1) high entry pressure to prevent the buoyancy driven upwards escape of CO2; (2) low permeability to minimize the upwards flux of brine displaced by the CO2; and (3) high strength to ensure that pressure build up does not cause caprock failure. We propose an alternative approach for cases when the above requirements are not met. The approach consists of extracting brine from the storage formation and then re-injecting it so that it mixes with CO2 at depth in the injection well. Mixing at depth reduces the pressure required for brine and CO2 at the surface. This CO2-saturated brine will sink to the aquifer bottom because it is denser than resident brine, which eliminates the risk of buoyant escape of CO2. The method is particularly favorable when the aquifer dips, because CO2-saturated brine will tend to flow downslope. We perform two- and three-dimensional numerical simulations to study how far upslope the extraction well needs to be located to ensure a very long operation without CO2 ever breaking through. Several sets of simulations were carried out to evaluate the effect of slope, temperature, pressure and CO2 concentration, which is significantly reduced if flue gas (i.e., without capture) is mixed with the brine. We analyze energy requirements to find that the system requires high permeability to be viable, but its performance is improved by taking advantage of the thermal energy of the extracted brine.