Chimeric antigen receptor T-cells (CAR-T) have emerged as an exciting new approach for the treatment of cancer, with over 500 CAR-T worldwide clinical trials underway in 2021.1 Conventional thinking was that all CAR-T cells have similar therapeutic potential, however new research has found that not all CAR-T cells are created equal. A recent study showed that CAR-T cells with lower levels of glycolysis led to better clinical responses in leukemia patients.2 Here we use VisionSort to identify metabolically distinct CAR-T cell subsets without the use of molecular labels or tags as an approach to identify cells with the potential for higher therapeutic efficacy.