Whether a vacancy is charged in a device or not depends on the location of the defect level in relation to the band edges. In an atomistic first-principles simulation we don't preassume any of that, you would just enter the vacancy as is, and then look for the charge around it in the converged state of the calculation.