Marine sponges are abundant and ecologically important components of coral reefs and have been shown to harbour exceptionally high microbial densities, which can differ substantially among sponge species. However, this dichotomy between high and low microbial abundance (HMA, LMA) sponges is still not fully understood, particularly as concerns the archaeal community. This study aims to fill this gap by analysing (using 454-pyrosequencing of the 16S rRNA gene) how the archaeal community varies among known LMA (Stylissa carteri, and Stylissa massa), known HMA (Hyrtios erectus and Xestospongia testudinaria) and unknown HMA/LMA status sponge species (Ectyoplasia coccinea, Paratetilla bacca and Petrosia aff. spheroida) collected in a remote location in which very few sponge microbial composition studies have been previously performed (Mayotte, Comores archipelago, France) and comparing the results with those reported in four other geographical areas. Based on archaeal community composition, the known LMA sponges formed a distinct cluster together with Paratetilla bacca, Ectyoplasia coccinea and seawater while the known HMA sponge X. testudinaria formed a cluster with Petrosia aff. spheroida. The known HMA sponge H. erectus, in turn, had an intermediate archaeal community between HMA sponges and sediment samples. In addition to the above, we also showed significant compositional congruence between archaeal and bacterial communities sampled from the same sponge individuals. HMA sponges were mainly dominated by members assigned to the genus Nitrosopumilus while LMA sponges were mainly dominated by members assigned to the genus Cenarchaeum. In general, there was no clear difference in richness between HMA and LMA sponges. Evenness, however, was higher in HMA than LMA sponges. Whilst the present study corroborates some of the traits commonly associated with the HMA–LMA dichotomy (higher evenness in Mayotte HMA sponges), this was not consistent across geographical areas showing that more research is needed to fully understand the HMA/LMA dichotomy as concerns Archaea.