Purpose of review In nephrogenic diabetes insipidus (NDI), the kidney is unable to concentrate urine despite elevated concentrations of the antidiuretic hormone arginine-vasopressin. In congenital NDI, polyuria and polydipsia are present from birth and should be immediately recognized to avoid severe episodes of dehydration. Unfortunately, NDI is still often recognized late after a ‘diagnostic odyssey’ involving false leads and dangerous treatments. Once diagnosed, appropriate treatment can be started. Moreover, laboratory studies have identified promising new compounds, which may help achieve urinary concentration independent of vasopressin. Recent findings MAGED2 mutations caused X-linked polyhydramnios with prematurity and a severe but transient form of antenatal Bartter's syndrome. We distinguish two types of hereditary NDI: a ‘pure’ type with loss of water only and a complex type with loss of water and ions. Mutations in the AVPR2 or AQP2 genes, encoding the vasopressin V2 receptor and the water channel Aquaporin2, respectively, lead to a ‘pure’ NDI with loss of water but normal conservation of ions. Mutations in genes that encode membrane proteins involved in sodium chloride reabsorption in the thick ascending limb of Henle's loop lead to Bartter syndrome, a complex polyuric-polydipsic disorder often presenting with polyhydramnios. A new variant of this was recently identified: seven families were described with transient antenatal Bartter's syndrome, polyhydramnios and MAGED2 mutations. Multiple compounds have been identified experimentally that may stimulate urinary concentration independently of the vasopressin V2 receptor. These compounds may provide new treatments for patients with X-linked NDI. Summary A plea for early consideration of the diagnosis of NDI, confirmation by phenotypic and/or genetic testing and appropriate adjustment of treatment in affected patients.