The often-used phrase ‘the uplift of the Tibetan Plateau’ implies a flat-surfaced Tibet rose as a coherent entity, and that uplift was driven entirely by the collision and northward movement of India. Here, we argue that these are misconceptions derived in large part from simplistic geodynamic and climate modeling, as well as proxy misinterpretation. The growth of Tibet was a complex process involving mostly Mesozoic collisions of several Gondwanan terranes with Asia, thickening the crust and generating complex relief before the arrival of India. In this review, Earth system modeling, paleoaltimetry proxies and fossil finds contribute to a new synthetic view of the topographic evolution of Tibet. A notable feature overlooked in previous models of plateau formation was the persistence through much of the Cenozoic of a wide east–west orientated deep central valley, and the formation of a plateau occurred only in the late Neogene through compression and internal sedimentation.

The uplift history of south-eastern Tibet is crucial to understanding processes driving the tectonic evolution of the Tibetan Plateau and surrounding areas. Underpinning existing palaeoaltimetric studies has been regional mapping based in large part on biostratigraphy that assumes a Neogene modernization of the highly diverse, but threatened, Asian biota. Here, with new radiometric dating and newly collected plant-fossil archives, we quantify the surface height of part of the south-eastern margin of Tibet in the latest Eocene (∼34 Ma) to be ∼3 km and rising, possibly attaining its present elevation (3.9 km) in the early Oligocene. We also find that the Eocene–Oligocene transition in south-eastern Tibet witnessed leaf-size diminution and a floral composition change from sub-tropical/warm temperate to cool temperate, likely reflective of both uplift and secular climate change, and that, by the latest Eocene, floral modernization on Tibet had already taken place, implying modernization was deeply rooted in the Palaeogene.