Myeloid leukemias are clonal hematopoietic stem cell disorders characterized either by proliferation of one or more of the myeloid lineages (chronic myelogenous leukemia) or by clonal expansion of myeloid blasts (acute myeloid leukemia). Over the past several years our knowledge of these hematologic malignancies has increased tremendously. The result is a classification that incorporates morphologic, immunophenotypic, genetic and clinical features in an attempt to define biologically and clinically relevant entities. Nevertheless, in many tumor subtypes the pathogenic event is still unknown. Furthermore, well-defined leukemia subgroups exhibit considerable heterogeneity, arousing the suspicion that several molecularly distinct subtypes might exist within the same cytogenetic category. Therefore, an ideal classification system would ultimately be based on the underlying molecular pathogenesis, but such knowledge is not yet available. However, by surveying the expression levels of thousands of genes in parallel, DNA microarrays have recently contributed to an increasingly refined molecular taxonomy of myeloid disorders. This powerful technology is becoming well established and has been used to diagnosis cancer and predict clinical outcome, to discover novel tumor subclasses, to gain insights into pathogenesis, and to identify new therapeutic targets. While many challenges remain ahead, genomic technologies have already demonstrated tremendous potential. We expect whole genome approaches will significantly contribute to a better understanding of the pathogenesis and result in a refined molecular classification of myeloid leukemias.