Stable isotope analysis has been widely used to aid the source identification of methane. However, the isotopic (13C/12C and D/H) and isotopologue (13CH3D and 12CH2D2) signatures of microbial methane in natural environments are often different from those in laboratory cultures in which methanogens are typically grown under optimal conditions. Growth phase and hydrogen (H2) concentration have been proposed as factors controlling the isotopic compositions of methane, but their effects on the relationship among carbon, hydrogen and doubly-substituted “clumped” isotopologue systems have not been assessed in a quantitative framework. Here we experimentally investigate the bulk (δ13C and δD) and clumped (∆13CH3D) isotopologue compositions of methane produced by hyperthermophilic hydrogenotrophic (CO2-reducing) methanogens using batch and fed-batch systems at different growth phases and H2 mixing ratios (Methanocaldococcus bathoardescens at 82 or 60 °C and on 80 or 25% H2; Methanothermobacter thermautotrophicus ∆H at 65 °C and on 20, 5 or 1.6% H2). We observed a large range (18 to 63‰) of carbon isotope fractionations, with larger values observed during later growth phase, consistent with previous observations. In contrast, hydrogen isotope fractionations remained relatively constant at –317 ± 25‰. Linear growth was observed for experiments with M. bathoardescens, suggesting that dissolution of gaseous H2 into liquid media became the rate limit as cell density increased. Accordingly, the low (and undersaturated) dissolved H2 concentrations can explain the increased carbon isotope fractionations during the later growth phase. The δD and Δ13CH3D values indicated departure from equilibrium throughout experiments. As the cell density increased and dissolved H2 decreased, Δ13CH3D decreased (further departure from equilibrium), contrary to expectations from previous models. Our isotopologue flow network model reproduced the observed trends when the last H-addition step is less reversible relative to the first three H-addition steps (up to CH3-CoM). In this differential reversibility model, carbon, hydrogen and clumped isotopologue fractionations are largely controlled by the reversibility of the first three H-addition steps under high H2 concentrations; the last H-addition step becomes important under low H2. The magnitude of depletion and decreasing trend in Δ13CH3D values were reproduced when a large (≥6‰) secondary clumped kinetic isotope effect was considered in the model. This study highlights the advantage of combined bulk and clumped isotope analyses and the importance of physiological factors (growth phase) and energy availability (dissolved H2 concentration) when using isotope analyses to aid the source identification of methane.