Abstract The Nubra-Shyok confluence in northern Ladakh is a key area for understanding the tectonic evolution of NW Himalaya and provides the basis for linking the geology of Pakistan to that of Tibet. The geology of the confluence area has been the subject of much speculation centred mainly on the existence of ophiolites and their regional significance. These ophiolites are thought to represent the eastward extension of the Shyok Suture Zone (SSZ), which separates the Dras island arc from the southern margin of Eurasia, and which was overprinted by movement along the Khalsar Thrust (often thought to represent the eastern continuation of the Main Karakoram Thrust). The geology of the area is relatively complex and the little information available has hampered regional geological correlations. The Khalsar Thrust (KT) and the dextral Karakoram Fault (KF), two regional tectonic features of NW Himalaya, merge at the confluence defining a triple point and three blocks: the Ladakh block to the south, the Saltoro block to the northwest, and the Karakoram block to the northeast. Close to the triple point, the KF changes strike and movement direction. Movement vector analysis of the triple point indicates that the KT and the two parts of the KF could have moved contemporaneously, and allows prediction of the movement vectors across the faults. The KT and KF shear preferentially volcano-sedimentary rocks of the Shyok and Nubra formations, respectively. Contrary to previous interpretations, these sheared rocks do not represent disrupted ophiolites. Regional tectonic reconstructions, however, require suturing between the Ladakh block and Eurasia and the strike of the SSZ in Baltistan suggests that the suture zone might crop out north of the KT, either along the southern slopes of the Saltoro Range or further north along the Saltoro valley. In the few outcrops of the Saltoro block we were able to visit, we found no evidence of ophiolitic rocks. Instead we found outcrops of the calc-alkaline Tirit batholith. Although our observations do not confirm the presence of the suture-related rocks in the southern Saltoro block, this possibility cannot be ruled out. Zircons from a sample of Tirit granite (U-Pb ion-microprobe age) yielded an age centred at 68 ± 1 Ma. The similar range of modal composition and age of the Tirit and Ladakh batholiths suggest that they are part of the same magmatic event. This result and a number of other observations indicate that the post–75 Ma geology of the Ladakh and Saltoro blocks is similar. Thus, if there is a suture zone in the southern Saltoro block, suturing must have occurred before 75 Ma, as concluded by others along the same tectonic boundary to the west in Pakistan. The KF represents a much younger terrane boundary, juxtaposing rocks of the Ladakh and Saltoro blocks to those of the Karakoram terrane. Rocks related to suturing of continents were not found along the KF. Karakoram leucogranites cropping out in the southern part of Karakoram terrane yielded a U-Pb zircon age centred at 15.0 ± 0.4 Ma (2σ). Because these leucogranites were not found south of the KF, this fault must have initiated after leucogranite intrusion and must therefore be younger than 15 Ma old. At the confluence the KF cuts across the regional rock sequence than can be followed from Kohistan into Baltistan and into the confluence area. Movement on the fault displaces the sequence by approximately 150 km to southeastern Tibet where the regional rock sequence can be regained.