Genome-wide association studies (GWAS) have led to the identification of more than 100 common, low-penetrance loci for cancer. At these loci, common genetic variants are associated with moderate increases in risk, typically <1.5-fold. Almost all loci lie in genomic regions not previously suspected to be involved in cancer. A plausible functional basis for a few loci, such as FGFR2 for breast cancer and MSMB for prostate cancer, has been elucidated, but the majority are not understood and suggest new mechanisms of carcinogenesis. Most loci are specific to a single cancer type, and are often subtype specific (e.g. ER-positive breast cancer). There are notable differences in the genetic architecture for different cancer types, with a greater contribution of common variants for prostate cancer. The clinical utility of variants to predict individual disease risk of disease is currently limited, but this may change as more variants are identified.