Chronic kidney disease (CKD) affects ~10% of the global population, with considerable ethnic differences in prevalence and aetiology. We assembled genome-wide association studies (GWAS) of estimated glomerular filtration rate (eGFR), a measure of kidney function that defines CKD, in 312,468 individuals from four ancestry groups. We identified 93 loci (20 novel), which were delineated to 127 distinct association signals. These signals were homogenous across ancestries, and were enriched for coding exons, kidney-specific histone modifications, and binding sites for HDAC2 and EZH2. Fine-mapping revealed 40 high-confidence variants driving eGFR associations, and highlighted potential causal genes with kidney cell-type specific expression in glomerulus, and proximal and distal nephron. Mendelian randomisation (MR) highlighted causal effects of eGFR on overall and cause-specific CKD, kidney stone formation and diastolic blood pressure (DBP). These results define novel molecular mechanisms and effector genes for eGFR, offering insight into downstream clinical outcomes and potential routes to CKD treatment development.