Our understanding of the molecular pathogenesis of osteoarthritis has evolved substantially over the past 20 years, but the study of osteoarthritis (OA) is much older. In the mid 19th century in-depth anatomical studies of the pathological changes in the joint were published, showing that multiple tissues were affected. These included splitting, fibrillation and loss of articular cartilage; thickening of the synovium and capsule; new bone formation (osteophytes); and subchondral bone sclerosis. Later, with the development of improved imaging, this extended to recognising changes in ligament insertions, the fat pads, and regions of oedema of the subchondral bone. Together these cause pain and loss of joint function. Many proposals have been put forward to account for why tissues change in OA. From epidemiological studies it is hard to deny that the most important factor in OA development is mechanical stress although there has been a reluctance by the research community to accept this as a unifying driver of OA joint biology. The advent of agnostic molecular profiling of OA patient tissues at scale, detailed clinical phenotyping and recognition of shared pathways across scientific disciplines has greatly accelerated discovery. Arguably a compelling new hypothesis for OA pathogenesis is emerging that aligns with the epidemiology, biomechanics, biology and genetic studies. This perspective represents a non-systematic, but evidence-based account of the role of mechanobiological pathways in OA pathogenesis, which have important implications for future treatments.