Osteoarthritis is a common disease resulting in significant disability without approved disease-modifying treatment (other than total joint replacement). Stem cell-based therapy is being actively explored for the repair of cartilage lesions in the treatment and prevention of osteoarthritis. Embryonic stem cells are a very attractive source as they address many of the limitations inherent in autologous stem cells, such as variability in function and limited expansion. Over the past 20 years, there has been widespread interest in differentiating ESC into mesenchymal stem cells and chondroprogenitors with successful in vitro, ex vivo, and early animal studies. However, to date, none have progressed to clinical trials. In this review, we compare and contrast the various approaches to differentiating ESC; and discuss the benefits and drawbacks of each approach. Approaches relying on spontaneous differentiation are simpler but not as efficient as more targeted approaches. Methods replicating developmental biology are more efficient and reproducible but involve many steps in a complicated process. The small-molecule approach, arguably, combines the advantages of the above two methods because of the relative efficiency, reproducibility, and simplicity. To better understand the reasons for lack of progression to clinical applications, we explore technical, scientific, clinical, and regulatory challenges that remain to be overcome to achieve success in clinical applications.