Dosage compensation in flies,‭ ‬in contrast to mammals,‭ ‬increases gene expression from the single X chromosome in males.‭ ‬To achieve this the dosage compensation complex‭ (‬DCC‭) ‬in‭ ‬Drosophila melanogaster‭ ‬specifically recognizes the X chromosome and acetylates the histone H4‭ ‬of active genes.‭ ‬A set of genomic regions known as high-affinity sites‭ (‬HAS‭) ‬had been identified as the X-specific binding locations of the dosage compensation complex.‭ ‬However,‭ ‬HAS are scattered throughout the X chromosome but no logic behind such disposition is currently known.‭ ‬Using Hi-C data we demonstrate that HAS are non-randomly distributed,‭ ‬appearing almost exclusively at the boundaries of topologically associating domains‭ (‬TADs‭)‬.‭ ‬To understand the significance of this localization we analyzed Hi-C and‭ ‬4C-seq contact data and found that TAD boundaries display more contacts to each other as well a to other chromosomal regions as compared to any other genomic region.‭ ‬These features are sex-independent as Hi-C and‭ ‬4C-seq contact maps remain comparable between male,‭ ‬female and DCC knockdowns.‭ ‬Further experiments allowed us to differentiate a two step process in which dosage compensation progresses:‭ ‬First,‭ ‬the DCC‭ ‬targets the X chromosome by binding HAS at TAD boundaries,‭ ‬then the DCC‭ ‬spreads‭ ‬from HAS to reach spatially nearby genes.‭ ‬Thus,‭ ‬once the DCC is bound to the chromosome X,‭ ‬it spreads,‭ ‬probably by diffusion,‭ ‬to nearby chromatin.‭ ‬Based on our Hi-C and‭ ‬4C-seq data we argue that TAD boundaries offer the best location from which spatial spreading can reach most chromatin,‭ ‬this could explain why the high-affinity sites are almost exclusively found at boundaries.