Cow's milk is often severely criticized as a cause of multiple health problems, including an increased risk of fractures. A close look at the scientific literature shows a striking contradiction. On the one hand, experimental studies of surrogate markers (e.g., bone turnover markers and bone mineral density [BMD]) usually indicate benefits from drinking cow's milk. On the other, the findings from epidemiological studies are conflicting and disconcerting. In all age groups, including children and postmenopausal women, consuming cow's milk, powdered milk supplements, or whey protein is associated with a slower bone turnover and unchanged or higher BMD values. These benefits are particularly marked in populations where calcium deficiency is prevalent, for instance in Asian countries. No interventional studies have addressed the fracture risk potentially associated with drinking cow's milk. The only available data come from epidemiological observational studies, whose results are conflicting, with a lower fracture risk in some cases and no difference or a higher risk in others. Several hypotheses have been offered to explain these findings, such as a deleterious effect of D-galactose, lactose intolerance, and acid overload. Epidemiological studies face many obstacles when seeking to detect effects of a single food, particularly the multiplicity of interactions among foods. Furthermore, reliable dietary intake data must be collected over prolonged periods, often long before the occurrence of a fracture, and defective recall may therefore introduce a major yet often unrecognized bias, particularly in populations where calcium deficiency is uncommon. To date, there is no conclusive evidence that we should modify our currently high level of consumption of cow's milk.