The deep ocean is thought to have played a crucial role in modulating atmospheric CO2 changes, and thus reconstructions of deep ocean conditions can place important constraints on the past global carbon cycle. Some previous studies suggested that foraminiferal U/Ca could be used to infer seawater carbonate chemistry changes, but others showed complications from diagenesis and temperature. A recent downcore study suggested that foraminiferal U/Mn may be used for sedimentary redox‐conditions, but no core‐top work has been done to investigate factors affecting U/Mn. We investigate controlling factors on U/Ca and U/Mn in two benthic foraminiferal species from 120 global core‐tops and three Atlantic sediment cores. Our core‐top data reveal no significant correlation between core‐top benthic U/Ca and carbonate system parameters. The lack of an influence of deep‐water [ CO32−] on U/Ca is further supported by our downcore results. Together, our data highlight complications to use benthic U/Ca for deep‐water carbonate chemistry reconstructions. Although no correlation is found between core‐top U/Mn and hydrographic data, high‐resolution U/Mn and U/Ca in core TNO57‐21 show similar patterns to authigenic U (aU) and vary in tandem with atmospheric CO2 on millennial timescales. Changes in U/Mn, U/Ca and aU in TNO57‐21 may reflect postdepositional diagenesis linked to sedimentary oxygen, which is controlled by subantarctic surface productivity and ventilation of deep South Atlantic in the past. We suggest that benthic U/Mn and U/Ca may be used as auxiliary indicators for past sedimentary redox‐conditions and along with other proxies could reflect deep‐water oxygenation.