We study the causal effect of the CARES Act on consumer expenditure during the Covid-19 pandemic with the Facteus data set. We use the transaction data at the Zip code level of over twelve million bank cards in eighteen central states to estimate the treatment effect of the stimulus. This question is identified because of the geographic and temporal variation in receiving the deposit across Zip codes. Our empirical findings suggest that in the first weekend receiving the stimulus, the overall spending increase by 53%. In particular, the expenditure on essential and financial categories increases by 62% and 76%. Unlike the tax rebates in 2001 and economic stimulus payments in 2008, people spend most of the payments on non-durable goods and services such as grocery, utility, money order, Etc. The estimated effect is substantial for districts with lower education, lower white ratio, and higher poverty, where poverty generates the most striking heterogeneity.