Abstract Background The objective of this study was to forecast the impact of COVID-19 vaccination in the United States (US) and China, two countries at different epidemic phases. Methods A mathematical model describing SARS-CoV-2 transmission and disease progression was used to investigate vaccine impact. Impact was assessed both for a vaccine that prevents infection (VES = 95%) and a vaccine that prevents only disease (VEP = 95%). Results For VES = 95% and gradual easing of restrictions, vaccination in the US reduced the peak incidence of infection, disease, and death by >55% and cumulative incidence by >32%, and in China by >77% and >65%, respectively. Nearly three vaccinations were needed to avert one infection in the US, but only one was needed in China. For VEP = 95%, benefits of vaccination were half those for VES = 95%. In both countries, the impact of vaccination was substantially enhanced with rapid scale-up, vaccine coverage >50%, and slower or no easing of restrictions, particularly in the US. Conclusions COVID-19 vaccination can flatten, delay, and/or prevent future epidemic waves. However, vaccine impact is destined to be heterogeneous across countries because of an underlying “epidemiologic inequity” that reduces benefits for countries already at high incidence, such as the US. Despite 95% efficacy, actual vaccine impact could be meager in such countries, if vaccine scale-up is slow, acceptance of the vaccine is poor, or restrictions are eased prematurely. One Sentence Summary Vaccine impact will be heterogeneous across countries disadvantaging countries at high incidence. This heterogeneity can be alleviated with rapid vaccination scale-up and limited easing of restrictions.