Generation of granitic melt is believed to occur predominantly by melting through the breakdown of hydrous minerals. However, melting due to the influx of H2O has been recognized in anatectic amphibolite facies tonalitic grey gneisses, metagreywackes and low‐P metapelites, and has consequently been proposed as an alternative mechanism for the generation of granitic melt. Melting induced by H2O addition is recognized from voluminous melt production at relatively low temperature, where hydrous minerals are stable and anhydrous minerals are preferentially consumed during melting. Mineral equilibrium modelling to determine the P–T conditions, melt volumes, melting reactions and viable H2O sources reveals that the process is not restricted to specific compositions or P–T conditions, although lower pressure and lithologies with a low hydrous mineral content are more favourable. Melting reactions in all lithologies primarily consume quartz and feldspars to yield 5–6 mol.% melt for each mol.% of H2O added. aH2normalO remains constant at ~0.70 to 0.77 during progressive melting as long as alkali feldspar is present. Once alkali feldspar is exhausted, plagioclase becomes the main reactant, producing more tonalitic melt compositions with gradually higher aH2normalO. Our results demonstrate that, at the site of melting, melting is driven by diffusion of H2O into the target rock along chemical potential gradients, rather than the advective flow of a mechanically distinct water‐rich fluid phase. Melting will initiate and proceed as long as a μH2normalO gradient exists between the H2O source and target lithology. Our calculations show that an ordinary magma, such as an I‐type magma with typical H2O content, has a μH2normalO high enough to be a viable H2O source, allowing diffusive H2O‐fluxed melting to produce melt proportions and fertility comparable to that of dehydration melting. However, high degrees of partial melting require a considerable amount of H2O, which necessitates a continuously advecting H2O source such as a magma conduit or melt‐bearing shear zone. A magmatic H2O source at emplacement level will undergo a similar amount of crystallization as the melt fraction produced in the target rock such that there will be no net melt production. Considering that shear‐zone hosted magma conduits are localized features, diffusive H2O‐fluxed melting is likely to only be viable in a small fraction of the anatectic orogenic crust. Although it may play an important role in locally raising melt volumes and modifying magma chemistry through mingling and hybridization, it does not appear to, of itself, be able to generate significant volumes of granitic melt.