Copy number alterations (CNAs), or gains and losses of DNA segments, are common somatic changes implicated in oncogenesis and genomic instability (GI). Akin to single base substitutions, the accumulation of CNAs within cancer genomes is often non-random and can be attributed to aberrant activity of specific DNA damage and repair processes. Studying the higher-order patterns of CNAs – especially in heterogenous, CNA-prevalent diseases such as breast cancer – can illuminate the types and determinants of GI as well as their clinical implications. These patterns can be characterized using pre-defined scores or, more recently, through de novo CNA signature extraction via matrix factorization. Applying these approaches to 2,763 breast cancer patients from The Cancer Genome Atlas (TCGA) and METABRIC, we found CNA scores and signatures associated with genomic architecture, driver genes, clinical features, and patient outcomes. Critical oncogenes, such as ERBB2, ESR1, and CCND1, are preferentially amplified through tandem duplication-like processes in a subtype-specific manner. We also identified CNA signatures that capture diploid and tetraploid samples with relatively stable (“CN Quiet”) genomes – the latter of which represents a unique subset of whole genome doubled tumors. Moreover, multiple signatures co-occurred with the loss of BRCA1/2 – genes critical for proficient homologous recombination (HR). Intriguingly, these signatures demonstrated differential activity between tumors lacking BRCA1 or BRCA2 – suggesting CNA-based GI patterns may vary depending on the underlying cause of HR deficiency. Lastly, we identified a signature that strongly associates with chromothripsis-like patterns and the presence of extrachromosomal DNA (ecDNA) – particularly in HER2+ cases. Overall, our findings demonstrate how deep interrogation of CNA-based GI patterns in breast cancer can elucidate mutagenic processes while simultaneously providing clinically-relevant insights.