Microbial ecosystems can be sustained by hydrogen gas (H-2)-producing water-rock interactions in the Earth's subsurface and at deep ocean vents(1-4). Current estimates of global H-2 production from the marine lithosphere by water-rock reactions (hydration) are in the range of 10(11) moles per year(5-9). Recent explorations of saline fracture waters in the Precambrian continental subsurface have identified environments as rich in H-2 as hydrothermal vents and seafloor-spreading centres(1,2) and have suggested a link between dissolved H-2 and the radiolytic dissociation of water(10,11). However, extrapolation of a regional H-2 flux based on the deep gold mines of the Witwatersrand basin in South Africa(11) yields a contribution of the Precambrian lithosphere to global H-2 production that was thought to be negligible (0.009 x 10(11) moles per year)(6). Here we present a global compilation of published and new H-2 concentration data obtained from Precambrian rocks and find that the H-2 production potential of the Precambrian continental lithosphere has been underestimated. We suggest that this can be explained by a lack of consideration of additional H-2-producing reactions, such as serpentinization, and the absence of appropriate scaling of H-2 measurements from these environments to account for the fact that Precambrian crust represents over 70 per cent of global continental crust surface area(12). If H-2 production via both radiolysis and hydration reactions is taken into account, our estimate of H-2 production rates from the Precambrian continental lithosphere of 0.36-2.27 x 10(11) moles per year is comparable to estimates from marine systems.