Government interest in using financial incentives (FIs) to stimulate health behaviors is increasing. The cost of longer-term FI intervention may be prohibitive, however. This study examined the impact of FI withdrawal on physical activity. A large 25-week quasi-experiment was conducted with users of a government-funded FI-based mobile health application. Users (n=584,760) from three Canadian provinces were included. Daily FIs were removed in Ontario (ON; intervention) but not British Columbia (BC; control) or Newfoundland and Labrador (NL; control). Simple linear regression models examined weekly mean daily step count after FI withdrawal and showed step count decreased in all provinces with the largest decrease observed in ON - but not to the extent expected (i.e., 198 and 274 fewer steps/day vs. BC and NL, respectively). Contrary to more than 50 years of psychology research, our findings provide rigorous quasi-experimental evidence that contemporary FIs may drive sustained health behavior change.