Bat laryngeal echolocation is considered as one of the most complex and diverse modes of auditory sensory perception in animals and its evolutionary history has been the cause of many scientific controversies in the past two decades. To date, the majority of scientific evidence supports that bats (Chiroptera) are divided into two subordinal groups: Yinpterochiroptera, containing the laryngeal echolocating superfamily Rhinolophidae as sister taxa to the non-laryngeal echolocating family Pteropodidae; and Yangochiroptera, containing all other laryngeal echolocating lineages. This topology has led to an unanswered question in mammalian biology: was laryngeal echolocation lost in the ancestral pteropodids or gained convergently in the echolocating bat lineages? To date, there is insufficient and conflicting evidence from fossil, genomic, morphological and phylogenomic data to resolve this question. We detail an ontogenetic study of fetal cochlear development from seven species of bats and five outgroup mammals and show that in early fetal development, all bats including the non-laryngeal echolocating pteropodids have a similarly large cochlea typically associated with laryngeal echolocation abilities. The subsequent cochlear growth rate in the pteropodids is the slowest of all mammals and leads to the pteropodids and the non-echolocating lineages eventually sharing a similar cochlear morphospace as adults. The results suggest that pteropodids maintain a vestigial developmental stage indicative of past echolocation capabilities and thus support a single origin of laryngeal echolocation in bats.