Using a randomized, crossover approach, cyclists (N = 20, overnight fasted) engaged in three 75-km time trials while ingesting water (WAT) or carbohydrate (0.2 g/kg every 15 minutes) from bananas (BAN) or a 6% sugar beverage (SUG). Blood samples were collected pre-exercise and 0 h-, 1.5 h-, and 21 h-post-exercise, and analyzed for NK cytotoxicity activity (NKCA) using pure NK cell populations. The two carbohydrate trials (BAN, SUG) compared to WAT were associated with higher post-exercise glucose, and lower cortisol, total blood leukocyte, neutrophil, and NK cell counts (interaction effects, P < 0.001). The immediate post-exercise increase in NK cell counts was higher in WAT (78%) compared to BAN (32%) and SUG (15%) trials (P ≤ 0.017). The 1.5 h post-exercise decrease in NK cell counts did not differ after WAT (−46%), BAN (−46%), and SUG (−51%) trials. The pattern of change in post-exercise NKCA differed between trials (P < 0.001). The 1.5 h post-exercise decreases in NKCA were 23%, 29%, and 33% in the WAT, BAN, and SUG trials, respectively, but trial contrasts did not differ significantly. Carbohydrate ingestion from BAN or SUG attenuated immediate-post-exercise increases in leukocyte, neutrophil, and NK cell counts, but did not counter the 1.5-h decreases in NK cell counts and NKCA.