Since the birth of the first prototype of the modern reaction turbine, cavitation as conjectured by Euler in 1754 always presents as a challenge. Following his theory, the evolution of modern reaction (Francis and Kaplan) turbines has been completed by adding the final piece of the element ‘draft-tube’ that enables turbines to explore water energy at efficiencies of almost 100%. However, during the last two and a half centuries, with increasing unit capacity and specific speed, the problem of cavitation has been manifested and complicated by the draft-tube surges rather than being solved. Particularly, during the last 20 years, the fierce competition in the international market for extremely large turbines with compact design has encouraged the development of giant Francis turbines of 700–1000 MW. The first group (24 units) of such giant turbines of 700 MW each was installed in the Three Gorges project. Immediately after commission, a strange erosion phenomenon appeared on the guide vane of the machines that has puzzled professionals. From a multi-disciplinary analysis, this Three Gorges puzzle could reflect an unknown type of cavitation inception presumably triggered by turbulence production from the boundary-layer streak transitional process. It thus presents a fresh challenge not only to this old turbine industry, but also to the fundamental sciences.