Identification of independent evolutionary lineages is a modern challenge for biologists especially in crops with high potential. Molecular methods, usually used to explore intraspecific variation, also offer important information regarding how and when lineages have evolved and spread. Today, several tools to define the demarcation of lineages based on single or multi-locus data are available, permitting objective results to be obtained. We tested nuclear and plastid loci using five different approaches to delimit the major lineages and estimate the divergence times in giant reed (Arundo donax L.). It is an emerging energy crop, widespread in sub-tropical and temperate regions but particularly aggressive outside its native range. Our results support the presence of six lineages distributed from the Asian regions to the Mediterranean basin and indicate the western and southern edges of the Qinghai-Tibet Plateau as the putative area of origin and source of ancient and cryptic lineages. We suppose it could be an area that has fertile populations. These results should be considered in breeding programmes to produce new cultivars, especially because the absence of fertility for Mediterranean giant reed lineages reduces the opportunity of crop improvement. Moreover, the description of multiple lineages and identifying its putative ancestral area may be helpful to researchers working on invasiveness management, phytoremediation programmes and conservation biology.