Background Adolescent cannabis use predicts the onset of later illicit drug use. In contrast, little is known about whether cannabis in young adulthood also predicts subsequent progression or cessation of licit or illicit drug use. Methods 13-year longitudinal cohort study with recruitment in secondary school students in Victoria, Australia. There were six waves of adolescent data collection (mean age 14.9-17.4 years) followed by three in young adulthood (mean age 20.7, 24.1 and 29.0 years). Discrete-time proportional hazards models were used to assess predictive associations between cannabis use frequency (occasional (<weekly), weekly to less than daily and daily) in 1756 participants in earlier young adult waves and subsequent cigarette smoking, high-risk alcohol use and amphetamine, ecstasy and cocaine use, including incident use (uptake) and cessation in later young adult waves. Results Compared with continuing occasional cannabis use: (1) never use provided the strongest protection from uptake of all drugs; (2) quitting cannabis lowered rates of illicit drug use uptake; (3) weekly+cannabis users had two to three times the rates of illicit drug use uptake, while daily users had six times the rate of uptake of cigarette smoking; and (4) never use of cannabis was associated with higher rates of cessation from licit drug use, while daily cannabis predicted lower cessation rates for all drugs except cocaine. Conclusions This study provides compelling evidence of the continuing association between cannabis, licit and other illicit drug use well into young adulthood. Preventing cannabis use uptake and use escalation remain crucial health aims given the burden associated with cigarette, alcohol and illicit drug use.