The abundance of methane isotopologues with two rare isotopes (e.g., 13CH3D) has been proposed as a tool to estimate the temperature at which methane is formed or thermally equilibrated. It has been shown, however, that microbial methane from surface environments and from laboratory cultures is characterized by low 13CH3D abundance, corresponding to anomalously high apparent 13CH3D equilibrium temperatures. We carried out a series of batch culture experiments to investigate the origin of the non-equilibrium signals in microbial methane by exploring a range of metabolic pathways, growth temperatures, and hydrogen isotope compositions of the media. We found that thermophilic methanogens (Methanocaldococcus jannaschii, Methanothermococcus thermolithotrophicus, and Methanocaldococcus bathoardescens) grown on H2 + CO2 at temperatures between 60 and 80 °C produced methane with Δ13CH3D values (defined as the deviation from stochastic abundance) of 0.5–2.5‰, corresponding to apparent 13CH3D equilibrium temperatures of 200–600 °C. Mesophilic methanogens (Methanosarcina barkeri and Methanosarcina mazei) grown on H2 + CO2, acetate, or methanol produced methane with consistently low Δ13CH3D values, down to −5.2‰. Closed system effects can explain part of the non-equilibrium signals for methane from thermophilic methanogens. Experiments with M. barkeri using D-spiked water or D-labeled acetate (CD3COO−) indicate that 1.6–1.9 out of four H atoms in methane originate from water, but Δ13CH3D values of product methane only weakly correlate with the D/H ratio of medium water. Our experimental results demonstrate that low Δ13CH3D values are not specific to the metabolic pathways of methanogenesis, suggesting that they could be produced during enzymatic reactions common in the three methanogenic pathways, such as the reduction of methyl-coenzyme M. Nonetheless CH bonds inherited from precursor methyl groups may also carry part of non-equilibrium signals.