INTRODUCTION: Household crowding deserves attention when evaluating the transmission intensity of SARS-CoV-2 in Brazil. We aimed to evaluate the association between household crowding and COVID-19 incidence. METHODS: Linear and Poisson regression analyses were used to assess the associations between indices of household crowding (high, average, low) and COVID-19 incidence estimates. RESULTS: Cities with a high index of household crowding were linked with a significantly higher COVID-19 incidence estimate (excess of 461 per 100,000; 95% confidence interval: 371-558 per 100,000). CONCLUSIONS: Crowding typically promotes virus transmission. Considering urban and housing structures is essential in designing mitigation strategies during a pandemic.