Diabetes remains a major burden. More than 200 million people are affected worldwide, which represents 6% of the population. The success achieved over the last decade with islet transplantation suggests that diabetes can be cured by the replenishment of deficient beta cells. These observations are proof-of-concept and have intensified interest in treating diabetes not only by cell transplantation but also by stem cells. Regeneration of beta cells from stem and progenitor cells is an attractive method to restore islet cell mass. Pancreatic stem/progenitor cells have been identified, and the formation of new beta cells from pancreatic duct, acinar and liver cells is an active area of investigation. Protocols for the in vitro differentiation of embryonic stem (ES) cells based on normal developmental processes have generated beta-like cells that produce high levels of insulin, even though at low efficiency and without full responsiveness to extracellular levels of glucose. Induced pluripotent stem (iPS) cells can also yield insulin-producing cells following similar approaches. Some agents including glucagon-like peptide-1/exendin-4 can stimulate the regeneration of beta cells in vivo as well as in vitro. Overexpression of embryonic transcription factors in stem cells could efficiently induce their differentiation into insulin-expressing cells. Recent progress in the search for new sources of beta cells has opened up several possibilities for the development of new treatments for diabetes.