No obviously inactivating variants are found in this gene, which provides at least negative data (humans generally seem to have at least somewhat functional ACE2). Exome sequencing across >200K humans has found a series of variants, mostly low-frequency, in the population.
Some of these variants may affect gene function, but it is not clear how, and none of them are obviously complete loss of function variants (e.g. nonsense mutations).
A more complete analysis of the data, including e.g. CNVs, might be necessary to fully address this question.