Intimate associations between different species drive community composition across ecosystems. Understanding the ecological and evolutionary drivers of these symbiotic associations is challenging because their structure eventually determines stability and resilience of the entire species network. Here, we compiled a detailed database on naturally occurring ant–symbiont networks in Europe to identify factors that affect symbiont network topology. These networks host an unrivalled diversity of macrosymbiotic associations, spanning the entire mutualism–antagonism continuum, including: (i) myrmecophiles – commensalistic and parasitic arthropods; (ii) trophobionts – mutualistic aphids, scale insects, planthoppers and caterpillars; (iii) social parasites – parasitic ant species; (iv) parasitic helminths; and (v) parasitic fungi. We dissected network topology to investigate what determines host specificity, symbiont species richness, and the capacity of different symbiont types to switch hosts. We found 722 macrosymbionts (multicellular symbionts) associated with European ants. Symbiont type explained host specificity and the average relatedness of the host species. Social parasites were associated with few hosts that were phylogenetically highly related, whereas the other symbiont types interacted with a larger number of hosts across a wider taxonomic distribution. The hosts of trophobionts were the least phylogenetically related across all symbiont types. Colony size, host range and habitat type predicted total symbiont richness: ant hosts with larger colony size, a larger distribution range or with a wider habitat range contained more symbiont species. However, we found that different sets of host factors affected diversity in the different types of symbionts. Ecological factors, such as colony size, host range and niche width predominantly determined myrmecophile species richness, whereas host phylogeny was the most important predictor of mutualistic trophobiont, social parasite and parasitic helminth species richness. Lastly, we found that hosts with a common biogeographic history support a more similar community of symbionts. Phylogenetically related hosts also shared more trophobionts, social parasites and helminths, but not myrmecophiles. Taken together, these results suggest that ecological and evolutionary processes structure host specificity and symbiont richness in large-scale ant–symbiont networks, but these drivers may shift in importance depending on the type of symbiosis. Our findings highlight the potential of well-characterized bipartite networks composed of different types of symbioses to identify candidate processes driving community composition.