Nosocomial outbreaks of vancomycin-resistant Enterococcus faecium (VREfm) are thought to occur by transmission of VREfm between patients, predicting that infection control interventions will limit cross-transmission. Despite implementation of such strategies, the incidence of VREfm infections continues to rise. We aimed to use genomics to better understand the epidemiology of E. faecium within a large hospital and investigate the reasons for failure of infection control strategies. Wholegenome sequencing was performed on 61 E. faecium (36 VREfm) isolates, predominately from blood cultures collected at a single hospital between 1998 and 2009, and on five vanB-positive anaerobic commensal bacteria isolated from human feces. Phylogenomic analysis and precise mapping of the vanB gene, which contains the Tn1549 transposon, showed that at least 18 of the 36 VREfm isolates had acquired the transposon via independent insertion events, indicating de novo generation of VREfm rather than cross-transmission. Furthermore, Tn1549 sequences found in 15 of the 36 VREfm isolates were the same as the Tn1549 sequence from one of the gut anaerobes. National and international comparator E. faecium isolates were phylogenetically interspersed with isolates from our hospital, suggesting that our findings might be globally representative. These data demonstrate that VREfm generation within a patient is common, presumably occurring in the human bowel during antibiotic therapy, and help explain our inability to reduce VREfm infections. A recommendation from our findings is that infection control practices should include screening patients for specific hospital clones of vancomycin-susceptible E. faecium rather than just VREfm. IMPORTANCE Enterococcus faecium is an increasingly important human pathogen causing predominantly antibiotic-resistant infections in hospitalized patients. Large amounts of health care funding are spent trying to control antibiotic-resistant bacteria in hospitals globally, yet in many institutions around the world, vancomycin-resistant E. faecium (VREfm) infections continue to rise. The new findings from this study help explain the failures of our current approaches to controlling vanB VREfm in health care institutions. Given the importance of this bacterium as a cause of hospital-acquired infections and the difficulties faced by infection control units in trying to prevent colonization in their institutions, the novel findings from this study provide evidence that a new approach to controlling VREfm in hospitals is required. In particular, more attention should be given to understanding the epidemiology of hospital-adapted vancomycin-susceptible E. faecium, and patients at higher risk for de novo generation of VREfm need to be identified and optimally managed.