Abstract The timing and extent of sulfide saturation have been suggested as controlling factors in the formation of economically significant porphyry Cu deposits in subduction zone settings. However, details on the sulfide saturation history in post-collisional porphyry systems remain ambiguous. Accordingly, we have characterized the whole-rock geochemistry, including platinum-group elements (PGE), of igneous intrusions in the post-collisional Chongjiang porphyry Cu–Mo–Au deposit (southern Tibet) and utilize this data in conjunction with zircon U–Pb geochronological results and sulfide chemistry to assess the timing of sulfide saturation, the nature and amount of magmatic sulfide produced. The Chongjiang intrusions (monzogranite, biotite monzogranite porphyry, granodiorite, dacite porphyry, and quartz diorite porphyry) and mafic microgranular enclaves (MMEs) have zircon U–Pb ages of 14.2 to 12.8 Ma. Covariations in whole-rock major and trace elements among the Chongjiang intrusions and MMEs, together with similarities in their Sr–Nd and zircon Hf isotope compositions, indicate that they are co-magmatic and crystallized from a juvenile lower crustal melt that mixed with mafic melt derived from the lithospheric mantle; this hybrid melt subsequently evolved via fractional crystallization. Trace-element ratios in zircon and temperature − ∆FMQ estimates of the different intrusions suggest that they all crystallized from oxidized (average ∆FMQ = 1.9–2.6) and water-rich magmas. Palladium contents and Pd/Pt ratios in the Chongjiang igneous intrusions increase with decreasing MgO up to 3.9 wt % MgO, after which they abruptly decrease. The initial increase in Pd/Pt ratios likely results from the fractionation of a Pt-rich mineral (e.g. Pt–Fe alloy). The decrease in Pd contents and Pd/Pt ratios at 3.9 wt % MgO likely results from sulfide saturation during magma evolution, but prior to volatile exsolution, which occurred at approximately 1.4 to 2.4 wt % MgO. The presence of magmatic sulfide inclusions in amphibole and magnetite in samples with 3.9 wt % MgO, and the geochemical compositions of sulfide inclusions suggest that they represented trapped sulfide liquid and intermediate solid solution. Results of Monte Carlo simulations demonstrate that 0.003 to 0.009 wt % magmatic sulfide is required to have fractionated from the magma to explain the decrease in Pd contents at 3.9 wt % MgO. Highly chalcophile elements, such as Pd, will be sequestered by the magmatic sulfide that saturates at depth, decreasing their concentrations in the residual silicate melt, whereas concentrations of the less chalcophile elements, such as Cu, Mo, and even Au, will not be as significantly affected. Consequently, sufficient concentrations of Cu–Mo–Au will remain in the residual melt and, upon reaching volatile saturation, can be transported by the vapor phase to form porphyry Cu–Mo–Au deposits. In the case of the Chongjiang deposit, sulfide saturation was likely triggered by the high pressures and/or depletion of FeO caused by the thick (~70 km) crust beneath the Gangdese belt. This contribution presents evidence of sulfide saturation in post-collisional magmatic systems, and demonstrates that the amount of magmatic sulfide produced is a critical factor in controlling the formation of post-collisional porphyry Cu deposits.