I think it is very likely that vaccinated people are contracting COVID-19 less than unvaccinated people, but because most testing is done on symptomatic people, and the vaccine is known to reduce symptoms, the sampling bias is likely overstating the effect.