For fifty years, heterogeneous photocatalysis has been considered as having potential to remove organic and microbiological pollutants from water under either artificial UV light or sunlight irradiation. However, after tens of thousands of published research papers, this system has been unable to move from laboratory bench scale to application mainly due to intrinsic thermodynamic limitations as well as kinetical and engineering drawbacks inherent to the heterogenous nature of the process undergoing to much longer treatment times when compared with other well-established water treatment options. Based on recent literature evidence, this critical review aims to describe and discuss the most significant drawbacks and limitations of TiO2-based photocatalytic processes to eliminate biological and chemical pollutants from water at a convenient degradation rate and in a noncomplex operational way. The properties and mechanistic action mode of TiO2 and similar semiconductors and their inherent engineering limitations for water treatment are highlighted. Moreover, the most common strategies unsuccessfully explored in the literature to overcome these drawbacks are described. Even if this analysis is specially focused on TiO2-based photocatalysis, most of the assertions and conclusions herein can be extrapolated to all the other semiconductors when the purpose is to reach water disinfection and detoxification. As 50 years ago, the main challenge remains to develop a new, modified (i.e., doped), or combined (i.e., junctions) photocatalyst generating high oxidative species at enough rates to significantly decrease the treatment time of this type of process (hours) in contrast with the tens of seconds or minutes observed for other systems.