BackgroundThe landscape of somatic mutations in prostate cancer (PCa) has quickly evolved over the past years.ResultsThis evolution was in part due to the improved quality and lower cost of genomic sequencing platforms available to an ever‐larger group of clinicians and researchers. The result of these efforts is a better understanding of early and late mutations that are enriched or nearly exclusive to treated PCa. There are, however, some important limitations to the current knowledge. The expanding variety of next‐generation sequencing (NGS) assays either capture a wide spectrum of mutations but at low coverage or are focused panels that cover a select number of genes, most often cancer‐related, at a deep coverage. Both of these approaches have their advantages, but ultimately miss low‐frequency mutations or fail to cover the spectrum of potential mutations. Additionally, some alterations, such as the common ETS gene fusions, require a mixture of DNA and RNA analysis to capture the true frequency. Finally, almost all studies rely on bulk PCa tumor samples, which fail to consider tumor heterogeneity. Given all these caveats, the true picture of the somatic landscape of PCa continues to develop.SummaryIn this review, the focus will be on how the landscape of mutations evolves during disease progression considering therapy. It will focus on a select group of early and late mutations and utilize SPOP mutations to illustrate recurrent alterations that may have clinical implications.