Chronic kidney diseases (CKDs) are currently classified according to their clinical features, associated comorbidities and pattern of injury on biopsy. Even within a given classification, considerable variation exists in disease presentation, progression and response to therapy, highlighting heterogeneity in the underlying biological mechanisms. As a result, patients and clinicians experience uncertainty when considering optimal treatment approaches and risk projection. Technological advances now enable large-scale datasets, including DNA and RNA sequence data, proteomics and metabolomics data, to be captured from individuals and groups of patients along the genotype-phenotype continuum of CKD. The ability to combine these high-dimensional datasets, in which the number of variables exceeds the number of clinical outcome observations, using computational approaches such as machine learning, provides an opportunity to re-classify patients into molecularly defined subgroups that better reflect underlying disease mechanisms. Patients with CKD are uniquely poised to benefit from these integrative, multi-omics approaches since the kidney biopsy, blood and urine samples used to generate these different types of molecular data are frequently obtained during routine clinical care. The ultimate goal of developing an integrated molecular classification is to improve diagnostic classification, risk stratification and assignment of molecular, disease-specific therapies to improve the care of patients with CKD.