A systematic review of ultra-processed food subtypes and risk of type 2 diabetes, combined with a cohort analysis, question the utility of broad food classifications for public health policy.