How do latent nuclear capabilities in the form of enrichment and reprocessing facilities affect interstate deterrence and coercion? Recent scholarship suggests that latency creates a “virtual” nuclear capability that can be used to deter threats or extract concessions, but these findings run counter to research demonstrating that states cannot deter without at least a deliverable nuclear device. This presentation aims to investigate this puzzling inconsistency by employing statistical analysis on an expanded dataset of latency measures and a variety of both military and bargaining outcomes.