For porphyry Cu deposits that formed during oceanic slab subduction, there is a general consensus that the ore-forming magmas evolved through fractionation of mafic magmas that were generated by slab fluid (± melt)–fluxed melting of the asthenospheric mantle wedge. This model, however, is not applicable to post-subduction porphyry Cu deposits because they formed distinctly after cessation of oceanic slab subduction. A popular model proposes that post-subduction porphyry Cu magmas were produced by partial melting of lower-crustal, sulfide-rich arc cumulates, with or without minor contributions from potassic mafic magmas. To reappraise the crustal melting model, we focused on one of the largest post-subduction porphyry Cu belts on Earth, which formed during the India-Asia collision in the Sanjiang region of southwestern China. Detailed petrographic studies and new Nd-Sr isotopic data from non-metasomatized versus metasomatized lower-crustal xenoliths suggest that previous models based on crustal melting rest upon wrong radiogenic isotope constraints due to pervasive metasomatism of the xenoliths. Based on traceelement quantitative modeling and regional-scale geochemical trends of magmatic rocks, we demonstrate that the Sanjiang post-subduction porphyry Cu magmas were produced by fractionation of potassic mafic magmas derived from lithospheric mantle. This study highlights that post-subduction porphyry Cu magmas attain their K-rich composition, and all the ore-forming ingredients, from subduction-modified lithospheric mantle, the existence of which may be a prerequisite for the formation of porphyry Cu deposits in post-subduction settings.