There is a compelling need to develop novel therapies for diabetes mellitus. Recent successes in the transplantation of islets of Langerhans are seen as a major breakthrough. However, there is huge disparity between potential recipients and the availability of donor tissue. Human embryonic stem cells induced to form pancreatic beta cells could provide a replenishable supply of tissue. Early studies on the spontaneous differentiation of mouse embryonic stem cells have laid the foundation for a more directed approach based on recapitulating the events that occur during the development of the pancreas in the mouse. A high yield of definitive endoderm has been achieved, and although beta-like cells can be generated in a step-wise manner, the efficiency is still low and the final product is not fully differentiated. Future challenges include generating fully functional islet cells under Xeno-free and chemically defined conditions and circumventing the need for immunosuppression.