Subduction is a key geodynamic feature on modern Earth that drives crustal chemical diversity, bridging the atmo-, hydro-, and lithosphere, but remains an enigmatic, unique planetary feature. Indisputable is the critical role of subduction in shaping Earth’s geomorphology and crustal dichotomy (ocean vs continental crust) and its impacts on long-term climate, making it arguably the most important process on present-day Earth across all geosciences. It is thus important to understand to what degree, or if at all, subduction was operational during the billions of years that led to our geological status quo.Here, we assess the feasibility of Archean subduction with a focus on early Earth geodynamics. We argue that convection-driven rifting, but not spreading, formed the first keels under the primordial crust, providing the necessary stability for crustal survival. These sections of crustal rejuvenation would counterintuitively forge the first stable proto-cratonic terranes, which later evolved into cratons. Hydrated upper crustal rocks were vital in generating early fluxed mantle melting and related volcanism, but also for partial melting in hydrated lower crustal sections within proto-cratons, giving rise to tonalite-trondhjemite granodiorites (TTGs). Both processes operated off- and on-craton, respectively, and required melting of hydrated crust and crustal convergence but are unrelated. Away from proto-cratonic regions of minor episodic divergence and rifting, relative motions were accommodated by convergence and shuffle tectonics, leading to Archean-style subduction in localised regions that were prone to destruction. This primitive form of subduction and crustal maturation has operated from the earliest Archean time in a plate-and-lid regime. Crucially, this ‘Archean subduction’ represents short-lived crustal shuffle-tectonics outside areas of today’s cratons with fluxed melting in upper mantle regions but does not resemble present-day Benioff-style subduction. The development of subduction akin to present-day processes towards the end of the Archean could plausibly have driven atmospheric oxygenation over a few hundred million years between ca. 2.8–2.3 Ga, with H-loss to space accompanied by atmospheric oxidation through subduction-related global volcanic SO2 emissions.