After primary infection with human cytomegalovirus (HCMV), which rarely causes any serious clinical problems in the immune competent, the virus persists subclinically for the lifetime of the host due, at least in part, to its ability to undergo latent infection. By contrast, HCMV can be a serious cause of morbidity, and in some cases mortality, upon primary infection of, or reactivation in, immune suppressed individuals. While current antivirals that target its lytic lifecycle have helped enormously in managing HCMV disease, to date, there are no available antivirals that target latent infection. In this review, we discuss research using natural and experimental models of latency that has led to some understanding of how HCMV latency is maintained, and reactivation controlled, in the myeloid lineage. Such analyses are now beginning to inform us of novel rationales that could allow the development of novel antivirals to target latency, itself.