The increasing availability of human pluripotent stem cells provides new prospects for neural-replacement strategies and disease-related basic research. With almost unlimited potential for self-renewal, the use of human embryonic stem cells (ESCs) bypasses the restricted supply and expandability of primary cells that has been a major bottleneck in previous neural transplantation approaches. Translation of developmental patterning and cell-type specification techniques to human ESC cultures enables in vitro generation of various neuronal and glial cell types. The derivation of stably proliferating neural stem cells from human ESCs further facilitates standardisation and circumvents the problem of batch-to-batch variations commonly encountered in "run-through" protocols, which promote terminal differentiation of pluripotent stem cells into somatic cell types without defined intermediate precursor stages. The advent of cell reprogramming offers an opportunity to translate these advances to induced pluripotent stem cells, thereby enabling the generation of neurons and glia from individual patients. Eventually, reprogramming could provide a supply of autologous neural cells for transplantation, and could lead to the establishment of cellular model systems of neurological diseases.