Misinformation becomes destructive when it involves a significant crowd. This group, with the ability to amplify disinformation, is commonly referred to as echo chambers. Detecting echo chambers is crucial as it can serve as an early warning for the emergence of disinformation circulation. Despite attracting researchers’ attention, their components and characteristics have not been thoroughly addressed. Because echo chambers form over time and may later disappear, this study aims to elucidate when echo chambers attain stability and an equilibrium of interactions among members. In this study, we employed a mixed methodology, incorporating both formal methods and a data-driven approach. Initially, we formally delineated the conditions under which echo chambers equilibrium holds. Subsequently, through experiments conducted on a large-scale dataset, we empirically demonstrated how these conditions manifest in practice. Experimental results on 22 million annotated records demonstrate that the aforementioned equilibrium can accurately detect echo chambers that have emerged in social media based on structure, belief purity, and reinforcement.