The purpose of this study was to define the proteomic and phosphoproteomic landscape of circulating extracellular vesicles (EVs) in people with normal glucose tolerance (NGT), prediabetes (PDM), and diabetes (T2DM). Archived serum samples from 30 human subjects (n = 10 per group, ORIGINS study, NCT02226640) were used. EVs were isolated using EVtrap®. Mass spectrometry-based methods were used to detect the global EV proteome and phosphoproteome. Differentially expressed features, correlation, enriched pathways, and enriched tissue-specific protein sets were identified using custom R scripts. Phosphosite-centric analyses were conducted using directPA and PhosR software packages. A total of 2372 unique EV proteins and 716 unique EV phosphoproteins were identified among all samples. Unsupervised clustering of the differentially expressed (fold change ≥ 2, p < 0.05, FDR < 0.05) proteins and, particularly, phosphoproteins showed excellent discrimination among the three groups. CDK1 and PKCδ appear to drive key upstream phosphorylation events that define the phosphoproteomic signatures of PDM and T2DM. Circulating EVs from people with diabetes carry increased levels of specific phosphorylated kinases (i.e., AKT1, GSK3B, LYN, MAP2K2, MYLK, and PRKCD) and could potentially distribute activated kinases systemically. Among characteristic changes in the PDM and T2DM EVs, “integrin switching” appeared to be a central feature. Proteins involved in oxidative phosphorylation (OXPHOS), known to be reduced in various tissues in diabetes, were significantly increased in EVs from PDM and T2DM, which suggests that an abnormally elevated EV-mediated secretion of OXPHOS components may underlie the development of diabetes. A highly enriched signature of liver-specific markers among the downregulated EV proteins and phosphoproteins in both PDM and T2DM groups was also detected. This suggests that an alteration in liver EV composition and/or secretion may occur early in prediabetes. This study identified EV proteomic and phosphoproteomic signatures in people with prediabetes and T2DM and provides novel insight into the pathobiology of diabetes.