Gang members are overrepresented among incarcerated populations in the United States. The link between incarceration and gang membership is beyond dispute, but serious questions remain about the causal mechanisms underlying this relationship. In this study, we develop and test theoretical models—origination, manifestation, and intensification—that focus on whether gang membership is exogenous or endogenous to incarceration. We used 7 years of monthly life calendar data nested within an 11‐wave longitudinal study of 1,336 serious offenders in Philadelphia and Phoenix to examine the effects of incarceration on gang membership. Multilevel models indicated that offenders were more likely to be in gangs while incarcerated in jail and prison settings than when not, although longer spells of incarceration corresponded with prolonged gang membership only in Phoenix. Incarceration in juvenile facilities maintained adverse between‐ and within‐individual effects on gang membership only in Phoenix. Additional descriptive findings revealed that gang status was durable to transitions into and out of incarcerated settings, and that more offenders exited than entered gangs while incarcerated. We situate these findings within our theoretical models and the body of knowledge on incarceration, concluding with a call for future research that is focused on the symbiosis between gangs in street and incarcerated settings.