The authors explore the critical issues remaining for the introduction of extreme ultraviolet lithography (EUVL) in semiconductor manufacturing. Among all technical issues, source power appears to be the most significant challenge that the technology is facing at this time. The lack of sufficiently high-power sources integrated in the first generation of full-field commercial scanners has profound implications on the remaining issues, and therefore on the risk associated with the insertion of EUVL technology. At the core of the problem is the cost of ownership of EUVL, which depends most heavily on source power. Moreover, the lifetime of the scanner mirrors and of the masks can only be tested properly with high-power sources on these first scanners. Without the ability to perform these tests, the technology might suffer unrecoverable delays. Fortunately, there has been good progress in the development of laser-produced plasma sources, which appear to be the most capable candidates for high-power scalability at this time. Furthermore, a booming new technology, fiber lasers, provides prospects of highly efficient and cost-effective EUV sources for high-volume manufacturing. Finally, significant progress has been made in the development of EUV resists to support the 32nm half-pitch node.