The upper reaches of the Indus River form a longitudinal network that drains a large portion of the western Himalaya. It plays an important role as a sediment routing system, feeding the Indus Delta and submarine fan, and has played a controlling role in the denudational history of the western Himalaya. Given the rivers long-term significance, its timing of initiation remains poorly constrained. The facies, palaeocurrents and provenance of the post-early Eocene Indus Group preserved along the upper reaches of the modern Indus reveal a history of fluvial/deltaic and deep-water sedimentation in an intermontane basin dominated by internal drainage. Hence, the Indus Group is not considered to represent the deposits of a palaeo-Indus River as previously thought. Illite crystallinity values and apatite fission track dating of the Indus Group suggest that the succession was buried to temperatures of 155–280°C, and that unroofing started in early Miocene times and proceeded at 0.1–0.4 mm a−1 to the present. The youngest sedimentary rocks preserved along the Indus Basin are early Miocene in age from the tectonostratigraphically equivalent deposits around Kargil to the west of the study area. The transition from sediment accumulation to erosional unroofing in early Miocene times coincides with accelerated regional unroofing of the High Himalayas to the south, and the initiation of the Indus Delta/submarine fan to the west. Differential uplift between the northward thrusting of the Zanskar and Indus sedimentary succession against the undeformed Ladakh Batholith provides a mechanism for post-early Miocene initiation of a major longitudinal river. Hence, early Miocene times is believed to represent the earliest possible age of initiation of the palaeo-Indus river following this course.