Sphingomonads comprise a physiologically versatile group within the Alphaproteobacteria that includes strains of interest for biotechnology, human health, and environmental nutrient cycling. In this study, we compared 26 sphingomonad genome sequences to gain insight into their ecology, metabolic versatility, and environmental adaptations. Our multilocus phylogenetic and average amino acid identity (AAI) analyses confirm that Sphingomonas, Sphingobium, Sphingopyxis, and Novosphingobium are well-resolved monophyletic groups with the exception of Sphingomonas sp. strain SKA58, which we propose belongs to the genus Sphingobium. Our pan-genomic analysis of sphingomonads reveals numerous species-specific open reading frames (ORFs) but few signatures of genus-specific cores. The organization and coding potential of the sphingomonad genomes appear to be highly variable, and plasmid-mediated gene transfer and chromosome-plasmid recombination, together with prophage- and transposon-mediated rearrangements, appear to play prominent roles in the genome evolution of this group. We find that many of the sphingomonad genomes encode numerous oxygenases and glycoside hydrolases, which are likely responsible for their ability to degrade various recalcitrant aromatic compounds and polysaccharides, respectively. Many of these enzymes are encoded on megaplasmids, suggesting that they may be readily transferred between species. We also identified enzymes putatively used for the catabolism of sulfonate and nitroaromatic compounds in many of the genomes, suggesting that plant-based compounds or chemical contaminants may be sources of nitrogen and sulfur. Many of these sphingomonads appear to be adapted to oligotrophic environments, but several contain genomic features indicative of host associations. Our work provides a basis for understanding the ecological strategies employed by sphingomonads and their role in environmental nutrient cycling.