Boiling water in small channels that are formed along turbine blades has been examined since the 1970s as a means to dissipating large amounts of heat. Later, similar geometries could be found in cooling systems for computers, fusion reactors, rocket nozzles, avionics, hybrid vehicle power electronics, and space systems. This paper addresses (a) the implementation of two-phase microchannel heat sinks in these applications, (b) the fluid physics and limitations of boiling in small passages, and effective tools for predicting the thermal performance of heat sinks, and (c) means to enhance this performance. It is shown that despite many hundreds of publications attempting to predict the performance of two-phase microchannel heat sinks, there are only a handful of predictive tools that can tackle broad ranges of geometrical and operating parameters or different fluids. Development of these tools is complicated by a lack of reliable databases and the drastic differences in boiling behavior of different fluids in small passages. For example, flow boiling of certain fluids in very small diameter channels may be no different than in macrochannels. Conversely, other fluids may exhibit considerable “confinement” even in seemingly large diameter channels. It is shown that cutting-edge heat transfer enhancement techniques, such as the use of nanofluids and carbon nanotube coatings, with proven merits to single-phase macrosystems, may not offer similar advantages to microchannel heat sinks. Better performance may be achieved by careful optimization of the heat sink’s geometrical parameters and by adapting a new class of hybrid cooling schemes that combine the benefits of microchannel flow with those of jet impingement.