Articular cartilage was long viewed as an inert, wear-prone bearing surface whose lifespan is spent by mechanical use. Contemporary evidence characterises articular cartilage as a mechanically regulated tissue in which physiological loading contributes to matrix homeostasis and suggests that osteoarthritis is better understood as a multifactorial disorder involving developmental, metabolic, and mechanotransductive dysregulation rather than simply as a consequence of mechanical wear. This narrative review synthesises converging strands of evidence on how exercise, physical activity, and mechanical loading affect knee articular cartilage, using the hip as a comparator. We outline the mechanobiology of the chondrocyte, in which mechanosensitive channels and downstream pathways translate load magnitude, mode, and duration into context-dependent biological responses, and we propose, as an integrative hypothesis rather than an established quantitative relationship, a physiological loading window that is multidimensional rather than a single dose axis. We review how compositional magnetic resonance imaging and molecular biomarkers make cartilage responses measurable in vivo, and we distinguish transient acute responses from chronic adaptation. We argue that load distribution matters as much as load magnitude, drawing on post-traumatic osteoarthritis. We appraise the clinical evidence in people at risk of, or with, knee osteoarthritis, where exercise is a first-line treatment that does not appear to harm cartilage, and we examine whether population-level running epidemiology, which is observational and susceptible to injury-related confounding and healthy-runner selection, is compatible with this model. Finally, we connect exercise mechanobiology to cartilage tissue engineering, where bioreactor conditioning and post-operative loading pose the same dose problem. We identify priority gaps: empirical testing of the loading-window hypothesis, a common metric of mechanical dose, the clinical translation of compositional imaging, and the under-representation of the hip.