The massive environmentally buffered nests of some social insects can contain millions of individuals and a wide variety of parasites, commensals and mutualists. We suggest that the ways in which these homeostatic fortress environments affect the evolution of social insect symbionts are relevant for epidemiology, evolutionary biology and macroecology. We contend that specialized parasites will tend to become less virulent and mutualists less cooperative, compared to those associated with solitary or small-colony hosts. These processes are expected to contribute to the very high symbiont diversity observed in these nests. We hypothesize that biodiversity gradients in these hotspots might be less affected by abiotic latitudinal clines than gradients in neighboring 'control' habitats. We suggest several research lines to test these ideas.