Streptococcus agalactiae (Group B streptococcus, GBS) is highly adapted to humans, where it is a normal constituent of the intestinal and vaginal flora. Yet, GBS has highly invasive potential and causes excessive inflammation, sepsis, and death at the beginning of life, in the elderly and in diabetic patients. Thus, GBS is a model pathobiont that thrives in the healthy host, but has not lost its potential virulence during coevolution with mankind. It remains incompletely understood how the innate immune system contains GBS in the natural niches, the intestinal and genital tracts, and which molecular events underlie breakdown of mucocutaneous resistance. Newborn infants between days 7 and 90 of life are at risk of a particularly striking sepsis manifestation (late-onset disease), where the transition from colonization to invasion and dissemination, and thus from health to severe sepsis is typically fulminant and not predictable. The great majority of late-onset sepsis cases are caused by one clone, GBS ST17, which expresses HvgA as a signature virulence factor and adhesin. In mice, HvgA promotes the crossing of both the mucosal and the blood–brain barrier. Expression levels of HvgA and other GBS virulence factors, such as pili and toxins, are regulated by the upstream two-component control system CovR/S. This in turn is modulated by acidic epithelial pH, high glucose levels, and during the passage through the mouse intestine. After invasion, GBS has the ability to subvert innate immunity by mechanisms like glycerinaldehyde-3-phosphate-dehydrogenase-dependent induction of IL-10 and β-protein binding to the inhibitory phagocyte receptors sialic acid binding immunoglobulin-like lectin 5 and 14. On the host side, sensing of GBS nucleic acids and lipopeptides by both Toll-like receptors and the inflammasome appears to be critical for host resistance against GBS. Yet, comprehensive models on the interplay between GBS and human immune cells at the colonizing site are just emerging.