We discovered at least six new paleosol occurrences within the Fortescue Group (Pilbara craton, Australia). Although largely reequilibrated by secondary processes, several of these paleosols preserve mineral assemblages inherited from early weathering, which raise questions about the interpretation of a reducing Late Archean atmosphere. This oxygenation process lasted long enough to trigger significant oxidative weathering as shown by numerous isotopic proxies in oceanic sediments. As a consequence, models for pCO2 estimations that rely on the anoxic atmosphere hypothesis are irrelevant, and solely the mass balance model calculations remain robust. Using climate modelling, we show that these estimates are consistent with an ice-free surface at the end of the Archean without the need of invoking a high pCH4. We propose that paleoproterozoic ice ages were triggered by the collapse of atmospheric methane that follows the oxygenation of the atmosphere, but also by a decrease of pCO2. For older points, we also show that the formation of clouds with large-sized dropplets is able to maintain an ice-free surface even at very low pCO2 (6 PAL). However, coupled climate-carbon simulations show that such a low pCO2 is unrealistic. In a general way, we argue that long-term pCO2 regulation is governed by the amount of continental surface exposed to weathering and their lithology.