The Paleocene Eocene Thermal Maximum (PETM) represents a major carbon cycle and climate perturbation that was associated with ocean de-oxygenation, in a qualitatively similar manner to the more extensive Mesozoic Oceanic Anoxic Events. Although indicators of ocean de-oxygenation are common for the PETM, and linked to biotic turnover, the global extent and temporal progression of de-oxygenation is poorly constrained. Here we present carbonate associated uranium isotope (δ238UCAU) data to reconstruct the evolution of global seawater δ238U, and hence quantify the expansion of anoxic U sinks on a global scale.This dataset contains trace element and U isotope data (238U/235U, expressed as δ238U) for the carbonate fraction of three well studied PETM sites; Site 865 (equatorial Pacific), Site 401 (Bay of Biscay) and Site 690 (Walvis Ridge). Samples are carbonate rich pelagic sediments, with a mixture of carbonate nannofossils, foraminifera and detrital clays. Bulk samples were selectively leached for the carbonate fraction using 1M ammonium acetate (pH 5) at room temperature for 24hrs. Trace element concentrations were measured on a Thermo-Finnigan Element XR and reported normalized to Ca. Uranium was purified by ion exchange chromatography and isotopes measured on a Thermo-Finnigan Neptune Plus. Uranium isotopes are reported as δ238U, where CRM-145 = 0‰.Sites 690 and 401 both show elevated U/Ca and δ238UCAU during the PETM and recovery interval, indicative of locally reducing conditions. By contrast, Site 865 records the global seawater δ238U and shows no resolvable change across the PETM. The lack of resolvable perturbation to the U-cycle during the event suggests a limited expansion of seafloor anoxia on a global scale. In the related publication we use this result, in conjunction with a biogeochemical model, to set an upper limit on the extent of global seafloor de-oxygenation. The model suggests that the new U isotope data, whilst also being consistent with plausible carbon emission scenarios and observations of carbon cycle recovery, permit a maximum ~10-fold expansion of anoxia to cover <2% of seafloor area.