Insulin regulates glucose transport through recruitment of GLUT4 to the plasma membrane (PM) where these transporters facilitate glucose uptake. The current model of GLUT4 recycling proposes a complex regulated system of GLUT4 cycling among specialized GLUT4 storage vesicles (GSV), intracellular compartments, and PM. Insulin has been variably reported to regulate GLUT4 recycling in several ways. However, despite this large number of processes apparently affected by insulin, recent experimental work by us and others strongly suggests that the main site of regulation of GLUT4 recycling and glucose uptake occurs at PM. To probe both GLUT4 organization in PM and its relationship to insulin-regulated recycling, we investigated GLUT4 dynamics in isolated rat adipose cells. We find: 1) clusters are generated by fusion with retention of GLUT4 in nascent domains; 2) GLUT4 is internalized at these domains after subsequent recruitment of clathrin, and 3) insulin induces a burst of GLUT4 exocytosis that mostly bypasses these domains and disperses GLUT4 directly into PM. The relationship between the spatial and temporal organization of plasma membrane (PM) glucose transporters is key to the regulation of cell metabolism. In addition to the complex recycling of GLUT4 among endosomal compartments, GLUT4-storage vesicles (GSV), and PM, we now know that their spatial organization in PM, where they facilitate glucose transport, depends upon insulin in a time-dependent fashion. The predominance in the basal state of PM GLUT4 cluster formation upon exocytosis of GSV gives way with insulin stimulation to a shift in GSV fusion to GLUT4 dispersal in PM. Further, GLUT4 clusters are found to nucleate clathrin assembly where most of the GLUT4 internalization from the cell surface then takes place. Thus, GLUT4 clusters represent a molecular organization mediating the transition between GLUT4 delivery and withdrawal from PM, depending upon insulin. While no molecular mechanism for GLUT4 clustering has been established, it was often attributed to accumulation of GLUT4 either in clathrin-coated pits or caveolae. Caveolar structures have been proposed to play some intermediate role in GLUT4 internalization. In our experiments we did not detect any significant co-localization of GLUT4 and caveolin in either basal or insulin-stimulated cells, which is consistent with another recent study arguing against a direct role for caveolae in GLUT4 recycling. While the role of clathrin in the recycling of GLUT4 is well supported, no evidence exists suggesting direct involvement of clathrin in GLUT4 clustering. We were able to separately measure the overall co-localization of GLUT4 and clathrin, as well as co-localization of surface-exposed GLUT4 with clathrin. Surprisingly, we found that surface-exposed GLUT4 have much less co-localization with clathrin than total GLUT4. Thus, a specific accumulation of GLUT4 in clathrin-coated pits is unlikely; it is more likely that these two molecules co-localize in intracellular compartments. Together with the fact that the majority of GLUT4 clusters exists at PM without showing any co-localization with clathrin, our data indicate that clathrin itself cannot account for the existence and formation of GLUT4 clusters at PM. The delivery of GLUT4 to PM through insulin-regulated exocytosis has been well documented by both biochemistry and live-cell imaging. Using TIRF microscopy, a number of groups detected single GSV fusion events using as a criterion for fusion the post-fusion dispersal of fluorescently labeled GLUT4-GFP. However, the number of fusion events measured microscopically was fewer than the number expected from biochemical and physiological approaches. Surprisingly, we found that many fusion events were not associated with dispersal of GLUT4 from the site of fusion, explaining why the number of fusion events detected only by GLUT4 dispersal was underestimated. Fusion-with-retention was predominant in the basal state where we observed that almost all fusion events detected by IRAP-pHluorin flash were not accompanied by dispersal of GLUT4 into PM. The result of this fusion-with-retention is the creation of de novo GLUT4 clusters at PM. Consistent with previous results, insulin increased the overall number of fusion events. Interestingly, insulin not only affected the number of fusion events, but also dramatically shifted the mode of fusion towards fusion with dispersal of GLUT4 into PM. This finding is also consistent with the pronounced increase of diffuse HA-antibody staining of PM and corresponding shift in the relative amount of GLUT4 from clusters to monomers. This observation further implies that while clustered and monomeric GLUT4 co-exist in a steady state, the relative amounts of GLUT4 in these pools can be differentially regulated by insulin. Interestingly, the insulin-stimulated increase of GLUT4 fusion was transient, and after a pronounced peak at 2-3 min, the fusion frequency declined to a level only slightly above the basal. While old models of GLUT4 recycling predict an over-all increase of GLUT4 recycling in response to insulin, our results fit the prediction of a quantum release model. However, while the older model considers insulin to regulate the size of the active pool available for GLUT4 recycling, our model includes all GLUT4 in the recycling process and addresses the existence of GLUT4 clusters as a distinct pool. One important feature of our model, based on the experimental observations, is the restriction of GLUT4 internalization to the clusters. Our data suggest that the major part of GLUT4 internalization occurs from clusters via recruitment of clathrin, while other clathrin-coated pits outside the clusters do not efficiently endocytose GLUT4. This organization of GLUT4 recycling provides flexibility to upregulate PM GLUT4 (in monomeric states) separately from GLUT4 available for endocytosis (clustered). Taken together, the data presented in this study suggest that GLUT4 clusters may function as intermediate hubs from the time of GLUT4 exocytosis until their internalization. In the basal state, these domains appear to play the major role in regulating the recycling of GLUT4 between PM and the intracellular pool of GSV. In the insulin-stimulated state, a rapid increase of PM GLUT4 is achieved by an increase in GSV fusion, particularly events with full and immediate release of GLUT4 molecules diffusely into PM. However, the rate of internalization and recycling of GLUT4 into GSV must now include a new parameter, the time it takes for GLUT4 to reach an uptake site; the rate-limiting step in this trafficking process remains to be determined. This kinetics, through GLUT4 hubs, must ultimately determine the new equilibrium GLUT4 activity level set by insulin stimulation. Thus, they are of particular interest in pathological states in which the relationship between insulin blood levels and GLUT4 activity is disrupted.