The formation of the Moon is thought to be the result of a giant impact between a Mercury-to-proto-Earth-sized body and the proto-Earth. However, the initial thermal state of the Moon following its accretion is not well constrained by geochemical data. Here, we provide geochemical evidence that supports a high-temperature origin of the Moon by performing high-temperature (1973–2873 K) metal-silicate partitioning experiments, simulating core formation in the newly-formed Moon. Results indicate that the observed lunar mantle depletions of Ni and Co record extreme temperatures (>2600–3700 K depending on assumptions about the composition of the lunar core) during lunar core formation. This temperature range is within range of the modeled silicate evaporation buffer in a synestia-type environment. Our results provide independent geochemical support for a giant-impact origin of the Moon and show that lunar thermal models should start with a fully molten Moon. Our results also provide quantitative constraints on the effects of high-temperature lunar differentiation on the lunar mantle geochemistry of volatile, and potentially siderophile elements Cu, Zn, Ga, Ge, Se, Sn, Cd, In, Te and Pb. At the extreme temperatures recorded by Ni and Co, many of these elements behave insufficiently siderophile to explain their depletions by core formation only, consistent with the inferred volatility-related loss of Cr, Cu, Zn, Ga and Sn during the Moon-forming event and/or subsequent magma-ocean degassing.@en