A growing number of hyper-partisan alternative media outlets have sprung up online to challenge mainstream journalism. However, research on news sharing in this particular media environment is lacking. Based on the virality of seventeen partisan outlets’ coverage of immigration and using the latest computational linguistic algorithm, the present study probes how hyper-partisan news sharing is related to source transparency, content styles, and moral framing. The study finds that the most shared articles reveal author names, but not necessarily other types of author information. The study uncovers a salient link between moral frames and virality. In particular, audiences are more sensitive to moral frames that emphasize authority/respect, fairness/reciprocity, and harm/care.