Energy-dense feeds commonly provided to equine athletes may be high in starch, which alter gastrointestinal microbiota and could promote systemic inflammation. To test the hypothesis that exercise-induced inflammation would be greater in horses receiving a starch- versus fiber-based concentrate. Quarter Horses (mean±SD 16±1mo; 337±30kg) received either a fiber-based control (CON; 7 fillies, 8 geldings) or an isocaloric, isonitrogenous starch concentrate (STARCH; 8 fillies, 7 geldings) for 24d. Fecal metagenomics were evaluated on d0 and 21. Blood inflammatory mediators were quantified on d0, d21, and surrounding a 2-h submaximal exercise test (SET) on d22. On d21, CON horses had greater Lactobacillaceae (∼5.7% vs. ∼2.4% in STARCH), while STARCH had greater Lachnospiraceae (∼38% vs. ∼32% in CON) but diet alone did not impact inflammatory markers. On d22, CRP increased at 24h post-SET in all horses (P<0.0001). By 48h, CRP returned to pre-SET in STARCH but remained elevated in CON (P=0.0005), resulting in greater CRP in CON than STARCH at 48h (P=0.02). TNFα increased from pre-SET to 1h in STARCH horses (P=0.02), then returned to pre by 6h. In CON horses, TNFα increased at 24h (P=0.04) and remained elevated at 48h (P=0.0005). Throughout the SET, CON had greater IL-10 than STARCH horses (P=0.005). SAA, IL-4, IL-8, and vascular endothelial growth factor (VEGF)-A were differentially impacted by the SET but were unaffected by diet. Contrary to our hypothesis, fiber-fed horses appeared to elicit a more robust acute inflammatory response to exercise than starch-fed horses despite an altered gastrointestinal microbiome.