Microglia contribute to Alzheimer’s Disease (AD) progression and are candidate therapeutic targets for disease modulation. Single cell transcriptomics demonstrate microglia adopt multiple phenotypes. However, identifying the human AD microglia profile was limited by small numbers. We employed fluorescence activated nuclei sorting prior to single-nucleus RNA-seq on human prefrontal cortices boosting the number of microglia analyzed per subject. We observed microglia subpopulations previously unrecognized in human brain and mapped their transcriptomic relationships by trajectory inference. We identified three clusters enriched for endolysosomal pathways, one of which showed differential expression of AD GWAS genes, in addition to genes implicated in nucleic acid detection and interferon signaling. We histologically demonstrated a microglia subpopulation with cytosolic nucleic acid and altered endolysosome morphology in AD tissue. Analysis of the “homeostatic” microglia cluster revealed a uniquely AD subpopulation. Our study demonstrates the value of deeply profiling microglia to explore the biological implications of microglia transcriptomic diversity.