Cannabis is the most commonly used illicit drug in psychosis patients and has been identified as a risk factor for relapse and subsequent hospital readmission, having substantial economic implications. To clarify the contribution of cannabis consumption to hospital readmission, a consecutive inpatient cohort of 161 early psychosis patients was included into the study. Data on cannabis use at admission and number of hospital readmissions and length of stay (LOS, number of inpatient days) in a 6-year follow-up was extracted from clinical notes. 62.4% of the patients had lifetime cannabis use. Their admission lasted on average 54.3 ± 75 days and over the following 6 years patients had 2.2 ± 2.8 hospital readmissions, for a total of 197.4 ± 331.5 days. Cannabis use significantly predicted the number of hospital readmissions and LOS in the following 6 years, the latter remaining significant after adjusting for use of other substance. Cannabis-using patients of male gender and Black ethnicity had a longer LOS at follow-up compared to female patients and other ethnic groups, respectively. Having a history of cannabis use when admitted to an early intervention inpatient unit for psychosis is associated with a higher number of subsequent hospital readmissions and a longer LOS, especially in male and Black patients.