Massively parallel DNA and RNA sequencing approaches have generated data on thousands of breast cancer genomes. In this review, we consider progress largely from the perspective of new concepts and hypotheses raised so far. These include challenges to the multistep model of breast carcinogenesis and the discovery of new defects in DNA repair through sequence analysis. Issues for functional genomics include the development of strategies to differentiate between mutations that are likely to drive carcinogenesis and bystander background mutations, as well as the importance of mechanistic studies that examine the role of mutations in genes with roles in splicing, histone methylation, and long non-coding RNA function. The application of genome-annotated patient-derived breast cancer xenografts as a potentially more reliable preclinical model is also discussed. Finally, we address the challenge of extracting medical value from genomic data. A weakness of many datasets is inadequate clinical annotation, which hampers the establishment of links between the mutation spectra and the efficacy of drugs or disease phenotypes. Tools such as dGene and the DGIdb are being developed to identify possible druggable mutations, but these programs are a work in progress since extensive molecular pharmacology is required to develop successful 'genome-forward' clinical trials. Examples are emerging, however, including targeting HER2 in HER2 mutant breast cancer and mutant ESR1 in ESR1 endocrine refractory luminal-type breast cancer. Finally, the integration of DNA- and RNA-based sequencing studies with mass spectrometry-based peptide sequencing and an unbiased determination of post-translational modifications promises a more complete view of the biochemistry of breast cancer cells and points toward a new discovery horizon in our understanding of the pathophysiology of this complex disease.