This paper is part of the special publication Continental transpressional and transtensional tectonics (eds R.E. Holdsworth, R.A. Strachan and J.F. Dewey). The Tibetan plateau north of the Himalaya has approximately double normal crustal thickness (60-75 km) and has been homogeneously shortened since the India-Asia collision at 60-50 Ma ago, yet, with minimal erosion rates, has almost no middle or deep crustal rocks exposed at the surface. In the Karakoram range, west of Tibet, early Tertiary crustal thickening and regional metamorphism resulted in Miocene crustal melting producing the Baltoro monzogranite-leucogranite batholith. Late Tertiary transpression along the western margin of the Tibetan plateau, caused by the continued northward penetration of India into Asia, led to exhumation of migmatites in the Pangong range and Baltoro-type granites along the Karakoram fault. Detailed studies of the Karakoram fault zone in northern Ladakh, India, show that Baltoro-type two-mica ± garnet leucogranites, intruded 21-18.0 Ma ago, have been offset a maximum of 150 km right- laterally. Average slip rates since 18.0±0.6 Ma (2σ) are 8.3 mm/a. 40Ar/ 39Ar mica cooling ages are 11.3 Ma on both sides of the main (southwestern) strand of the fault, suggesting that most of the exhumation of the Pangong migmatites and leucogranites must have occurred between 18.0 and 11.3 Ma. During this time, at slip rates of 8.3 mm/a the rocks would have moved horizontally right-laterally for c.56 km and been exhumed by c.20 km vertically during transpression, using the measured 20°plunge of lineations. The high exhumation rate (3.0 mm/a) and amount of erosion (20 km) inferred between 18.0 and 11.3 Ma may also reflect the partitioning between an early transpressional strain associated with crustal thickening and exhumation of the Pangong deep crustal migmatites and leucogranites, and a later dominantly dextral strike-slip phase of fault motion along the central part of the Karakoram fault from c.11 to 0 Ma. This timing may also coincide with the initiation of the N-S aligned normal faults and E-W extension in southern Tibet. We suggest that the relatively minor dextral offset (150 km) and the young age of initiation on this bounding fault do not support the model of large-scale extrusiion of Tibetan crust, but they suggest instead that deformation of Tibet was taken up predominantly by crustal thickening.