For a long time prokaryotic species definition has been under debate and a constant source of turmoil in microbiology. This has recently prompted the ASM to call for a scalable and reproducible technique, which uses meaningful commonalities to cluster microorganisms into groups corresponding to prokaryotic species. Whole-genome Average Nucleotide Identity (gANI) was previously suggested as a measure of genetic distance that generally agrees with prokaryotic species assignments based on the accepted best practices (DNA-DNA hybridization and 16S rDNA similarity). In this work, we prove that gANI is indeed the meaningful commonality based on which microorganisms can be grouped into the aforementioned clusters. By analyzing 1.76 million pairs of genomes we find that identification of the closest relatives of an organism via gANI is precise, scalable, reproducible, and reflects the evolutionary dynamics of microbes. We model the previously unexplored statistical properties of gANI using 6,000 microbial genomes and apply species-specific gANI cutoffs to reveal anomalies in the current taxonomic species definitions for almost 50percent of the species with multiple genome sequences. We also provide evidence of speciation events and genetic continuums in 17.8percent of those species. We consider disagreements between gANI-based groupings and named species and demonstrate that the former have all the desired features to serve as the much-needed natural groups for moving forward with taxonomy. Further, the groupings identified are presented in detail at http://ani.jgi-psf.org to facilitate comprehensive downstream analysis for researchers across different disciplines