Human pancreatic islets, which maintain glucose homeostasis via coordinate action of their constituent endocrine and affiliate cell types, are central to both type 2 diabetes (T2D) genetics and pathophysiology. Knowledge of islet cell type-specific alterations in T2D is lacking due to small sample numbers or limited single cell transcriptomes profiled per sample for comparison. Here, we present a comprehensive single cell transcriptome atlas of 245,878 human islet cells from a distinct 48-donor cohort spanning non-diabetic (ND), pre-diabetic, and T2D states. Unsupervised clustering of donor transcriptomes classified 14 cell types and indicated a 13% reduction in b-cells in T2D vs. ND islets, accompanied by reciprocal 10.5% declines in ‘insulin-secreting’ and 12.3% increases in ‘senescent’ b-cell subpopulations. Pseudo-bulk analyses across glycemic states revealed negligible gene expression changes in most cell types, except for b-cells. We identified 511 differentially expressed genes (DEGs) in T2D b-cells, 66.5% of which are novel and implicated in neurotransmission. Overlap with reported islet eQTLs identified 29 DEGs with aligned T2D genetic risk and subsequent gene expression programming in b-cells, meriting high-priority for therapeutic interventions. Additionally, we found 20 DEGs with divergent effects of T2D predisposition on gene expression among whole islets and b-cells. The generated transcriptomic data is provided as a free resource for islet biology research. Overall, our study develops a thorough understanding of T2D-regulated islet structural and functional integrity and, uncovers T2D effector genes with reinforced genetic and molecular impacts.