Résumé: In French Polynesia, the production and exportation of black pearls through the aquaculture of the black-lip pearl oyster Pinctada margaritifera provide the second largest economic income for the country after tourism. This industry entirely relies on the collection of natural spats from few highly recruiting lagoons. In recent years, pearl oyster producers have experienced variable success rates in spat collection, with significant spatial and temporal variability in spat supply, driving uncertainty in the future of pearl production. This study combines, for the first time in a farmed lagoon, genetic (SNPs), demographic (sex ratio, age), and biophysical data (larval dispersal modelling) to shed new light on population dynamics, connectivity, and spat recruitment in Ahe Atoll, a well-studied pearl farming site. Our results indicate that the geographical structuring of the natural populations and the contribution of both natural and exploited stocks to the production of spats result from the interaction of hydrodynamic features, life history traits and demographic parameters: the northeastern natural populations are older, not well connected to the southwestern natural populations and are not replenished by larvae produced by adjacent exploited populations. Moreover, we observe that the exploited populations did not contribute to larval production during our experiment, despite a sampling period set during the most productive season for spat collection. This is likely the result of a strong male bias in the exploited populations, coupled with a sweepstakes reproductive strategy of the species. Our results warrant further investigations over the future of the northeastern older natural populations and a reflection on the current perliculture techniques.