Abstract Branched glycerol dialkyl glycerol tetraether (brGDGT) paleothermometry is an important tool for paleoclimate research. Oxygen availability has been suggested to skew brGDGT‐inferred temperatures, but the effect is poorly constrained. Here, we examine the influence of dissolved oxygen (DO) on brGDGTs in lacustrine water filtrates. By examining three subgroups of samples collected within narrow temperature ranges, we more effectively isolate the influence of DO. We find significant correlations between DO and the MBT′5Me temperature index in all three constant‐temperature subgroups. Underlying shifts in key brGDGT abundances indicate the problem is likely to manifest in virtually all existing calibrations. The effects appear at even modest DO depletion, where existing anoxia biomarkers are unlikely to appear. Though creating a quantitative correction is not straightforward, we estimate maximum DO‐driven temperature offsets. These observations highlight a large and hitherto unquantified source of uncertainty in brGDGT‐based paleotemperature reconstructions and underscore the need for further research.