Social and economic disparities between racial/ethnic groups are a feature of the American context into which immigrants are incorporated, and a key determinant of population health. We ask whether racial/ethnic disparities in diabetes vary by nativity and whether native-immigrant disparities in diabetes vary by race and over time in the United States. Using the 2000-2015 National Health Interview Survey, we estimate logistic regressions to examine the interaction of race/ethnicity, nativity, and duration in the U.S. in shaping diabetes patterns. Relative to their native-born co-ethnics, foreign-born Asian adults experience a significant diabetes disadvantage, while foreign-born Hispanic, Black, and White adults experience a significant advantage. Adjusting for obesity, education, and other covariates eliminates the foreign-born advantage for Black and White adults, but it persists for Hispanic adults. The same adjustment accentuates the disadvantage for foreign-born Asian adults. For Black and Hispanic adults, the protective foreign-born effect erodes as duration in the U.S. increases. For foreign-born Asian adults, the immigrant disadvantage appears to grow with duration in the U.S. Relative to native-born White adults, all non-white groups regardless of nativity see a diabetes disadvantage because the racial/ethnic disadvantage either countervails a foreign-born advantage or amplifies a foreign-born disadvantage. Racial/ethnic differentials in diabetes are considerable, and are influenced by each group's nativity composition. Obesity and (for the foreign-born) time in the U.S. influence these disparities, but do not explain them. These findings underscore the importance of unmeasured, systemic determinants of health in America's race-conscious society.