The highly pathogenic H5N1 avian influenza A virus (IAV) clade 2.3.4.4b has spread globally and spilled over into multiple mammalian species, raising concerns about its pandemic potential. In late 2022, clade 2.3.4.4b viruses devastated seabird and marine mammal populations along the Pacific coast of South America. Here, we report the first evidence of H5 IAV infections in wild bats globally, focusing on common vampire bats (Desmodus rotundus) in coastal areas of Peru. Longitudinal serological screening, stable isotope analysis and metabarcoding revealed repeated exposures to H5 IAVs in vampire bats which feed on coastal wildlife species heavily impacted by the 2.3.4.4b epizootic, but no evidence of infection in populations without access to marine prey. We further report bat gene flow between IAV-exposed and IAV-naive populations, and IAV infections in a vampire bat colony that fed on both marine and terrestrial livestock prey, providing insights into how future IAV epizootics might spread spatially within bats and between marine and terrestrial ecosystems if a bat reservoir were established. Immunohistochemistry demonstrated that the H5 haemagglutinin protein binds to the upper respiratory tract of vampire bats, suggesting bat tissue susceptibility to H5 IAVs. Finally, vampire bat-derived kidney, liver, and lung cells supported entry, replication, and egress of avian and mammalian 2.3.4.4b viruses, confirming cellular infectivity. These results illustrate how combining ecological inference and experimental virology can pinpoint the species origins and biological significance of viral spillover at species interfaces. Recurrent exposures from marine wildlife, tissue and cellular susceptibility to H5N1 IAVs, and connections to other IAV-susceptible terrestrial mammals establish the prerequisite conditions for vampire bats to spread IAVs between marine and terrestrial environments or to form a novel reservoir of highly pathogenic IAVs.