Marine ammonia-oxidizing archaea of the phylum Thaumarchaeota are a cosmopolitan group of microorganisms representing a major fraction of the picoplankton in the ocean. The cytoplasmic membranes of Thaumarchaeota consist predominantly of intact polar isoprenoid glycerol dibiphytanyl glycerol tetraether (GDGT) lipids, which may be used as biomarkers for living Thaumarchaeota. Fossil thaumarchaeal GDGT core lipids accumulate in marine sediments and serve as the basis for geochemical proxies such as the TEX86 paleothermometer. Here, we demonstrate that the responses of membrane lipid compositions and resulting TEX86 values to growth temperature strongly diverge in three closely related thaumarchaeal pure cultures, i.e., Nitrosopumilus maritimus and two novel strains isolated from South Atlantic surface water, although the inventories of intact polar lipids and core lipids were overall similar in the three strains. N. maritimus and its closely related strain NAOA6 showed linear relationships of TEX86 and growth temperature but no correlation of TEX86 and temperature was observed in the more distantly related strain NAOA2. In contrast, the weighted average number of cycloalkyl moieties (ring index) was linearly correlated with growth temperature in all strains. This disparate relationship of TEX86 to growth temperature among closely related Thaumarchaeota suggests that the ring index but not the TEX86 ratio represents a universal response to growth temperature in marine planktonic Thaumarchaeota. Furthermore, the distinct TEX86-temperature relationships in the cultivated strains indicate that environmental GDGT signals may include an ecological component, which has important implications for ocean temperature reconstructions using the TEX86 proxy. In contrast, different growth medium salinities in the range 27–51‰ tested for N. maritimus showed no systematic effect on intact polar GDGT composition and TEX86. Similarly, N. maritimus showed only small changes in intact polar GDGT composition and TEX86 when grown at different medium pH in the range 7.3–7.9. Overall, our pure culture studies suggest that the TEX86 paleotemperature proxy is not solely dependent on growth temperature, but may amalgamate physiological, environmental, and ecological factors.