not-yet-known not-yet-known not-yet-known unknown Objective: Maternal genotypes may be useful to customize fetal growth assessment, but generalizability across diverse racial and ancestral groups remains uncertain. We assessed the generalizability of a genetic risk score for birth weight (GRS BW), derived from European ancestry participants, within a diverse U.S. cohort. Design: Secondary analysis of a prospective observational cohort of nulliparous patients. Setting: Eight U.S. recruitment centers. Population or Sample: Participants in the parent study with available maternal DNA. Methods: We used log-linear modeling to test the association of maternal GRS BW with fetal weight. We then assessed the robustness of the association by self-identified race and genetically predicted continental ancestry (GPA) groups. Main Outcome Measures: Association between GRS BW and fetal weight. Results: Among 8,147 eligible participants, GRS BW was associated with fetal weight (p<0.001). Across self-identified racial groups, the association was significant in White (p=0.007) and multiracial (p=0.03) groups but not in Black, Asian, or unknown race groups (p>0.09 for all). Among GPA groups, the association was significant among European (p=0.001) and American (p=0.02) ancestry groups but not African, East or South Asian, or unknown ancestry (p>0.05 for all). Conclusions: This GRS BW is not generalizable across races, highlighting the need for globally representative genetic discovery cohorts.