Context. Hi Intensity Mapping (IM) will be used to do precision cosmology, using many existing and upcoming radio observatories. It will measure the integrated Hi 21 cm emission signal from 'voxels' of the sky at different redshifts. The signal will be contaminated due to absorption, the largest component of which will be the flux absorbed by the Hi emitting sources themselves from the potentially bright flux incident on them from background radio continuum sources. Aims. We, for the first time, provide a quantitative estimate of the magnitude of the absorbed flux compared to the emitted Hi flux. The ratio of the two fluxes was calculated for various voxels placed at redshifts between 0.1 and 2.5. Methods. We used a cosmological sky simulation of the atomic Hi emission line, and summed over the emitted and absorbed fluxes for all sources within voxels at different redshift. In order to determine the absorbed flux, for each Hi source the flux incident from background radio continuum sources was estimated by determining the numbers, sizes, and redshift distribution of radio continuum sources that lie behind it, based on existing observations and simulations. The amount of this incident flux that is absorbed by each Hi source was calculated using a relation between integrated optical depth with Hi column density determined using observations of damped Lyman-α systems (DLAs) and sub-DLAs. Results. We find that for the same co-moving volume of sky, the Hi emission decreases quickly with increasing redshift, while the absorption varies much less with redshift and follows the redshift distribution of faint sources that dominate the number counts of radio continuum sources. This results in the fraction of absorption compared to emission to be negligible in the nearby Universe (up to a redshift of ∼0.5), increases to about 10% at a redshift of one, and continues to increase to about 30% up to a redshift of 2.5. These numbers can vary significantly due to the uncertainty on the exact form of the following relations: firstly, the number counts of radio continuum sources at sub-mJy flux densities; secondly, the relation between integrated optical depth and Hi column density of Hi sources; and thirdly, the redshift distribution of radio continuum sources up to the highest redshifts. Conclusions. Absorption of the flux incident from background radio continuum sources might become an important contaminant to Hi IM signals beyond redshifts of 0.5. The impact of absorption needs to be quantified more accurately using inputs from upcoming deep surveys of radio continuum sources, Hi absorption, and Hi emission with the Square Kilometre Array and its precursors.