Although diabetes is the most common cause of end-stage renal disease (ESRD) worldwide, most people with diabetic nephropathy will never develop ESRD but will instead die of cardiovascular (CV) disease (CVD). The first evidence of kidney injury in diabetes is often microalbuminuria, itself also an independent risk marker for CVD. Although the two processes are closely associated, the recent failure of antialbuminuric therapies to affect CV outcomes has encouraged a reconsideration of how albuminuria may occur in diabetes and how increased urinary albumin excretion may be indicative of CV risk. The relationship between CVD and urinary albumin content (even within the normal range) is widely considered to reflect the common underlying pathology of endothelial dysfunction. At the same time, recent years have witnessed a growing appreciation that diabetic albuminuria commonly arises from damage to glomerular podocytes, specialized epithelial cells acting as the final barrier to macromolecular flow into the urinary filtrate. These superficially discordant paradigms can be assimilated by the emerging concept of endothelial-podocyte crosstalk across the glomerular filtration barrier, whereby the actions of one type of cell may profoundly influence the function of the other. The bidirectional nature of this paracrine network is illustrated by the actions of the vascular endothelial growth factor-A (VEGF-A)/VEGF receptor-2 and activated protein C systems, among others. Identification of novel mediators of endothelial-podocyte crosstalk may lead to the development of more effective treatments for diabetic nephropathy and its sequelae.