Some per- and polyfluoroalkyl substances (PFAS) are classified as carcinogenic or possibly carcinogenic to humans by the World Health Organization, and there are cancer-related concerns in communities with known drinking water PFAS contamination. We conducted a ZIP Code Tabulation Area (ZCTA) level ecological study comparing cancer incidence rates in communities with and without PFAS-contaminated drinking water. We merged data on drinking water measurements for six PFAS from the U.S. Environmental Protection Agency’s Third Unregulated Contaminant Monitoring Rule (2013–2015), incident adult and child/adolescent cancer counts from six state cancer registries from 2008–2016, and 2010 census data. We assessed associations between ZCTA-level cancer incidence rates (31 adult and 18 child/adolescent cancer groups) and drinking water PFAS detection using Poisson regression models controlling for state, age group, sex, race/ethnicity, and year. We ran models with a random effect for ZCTA to account for clustering, and models without ZCTA random effects. We observed positive associations between any PFAS detection and incidence rates of adult colorectal [rate ratio (RR)=1.045 (95% confidence interval: 1.019–1.071)], kidney [RR=1.058 (1.022–1.095)], and cervical [RR=1.085 (1.018–1.158)] cancers (RRs from models with ZCTA random effects), and incidence rates of child/adolescent “germ cell tumors (GCTs), trophoblastic tumors, and neoplasms of gonads” [RR=1.104 (1.025–1.190)] and “malignant gonadal GCTs” [RR=1.113 (1.033–1.199)], which included child/adolescent testicular cancers (RRs from models without ZCTA random effects). The ZCTA-level associations we observed should be interpreted in the context of the ecological study design. Our ecological findings for kidney, colorectal, and germ cell cancers may corroborate other epidemiological or in vitro research, though data are sparse. Future research would benefit from individual-level data on exposures to specific PFAS types.