Abstract Introduction Cancer cachexia is associated with poor surgical outcomes, including increased length of stay, morbidity and mortality. Cachexia is a complex multi-faceted pathological process consistently associated with energy imbalance via hypermetabolism. Beyond this, pre-clinical animal models suggest mitochondrial dysfunction to be an important component of cachexia, although this is unconfirmed in humans. This study aimed to investigate the mitochondrial function of pre-surgical cancer patients, and the impact of exercise prehabilitation on this. Methods This study was approved by an NHS research ethics committee (IRAS: 275264). Male patients awaiting surgery with curative intent for colorectal and prostate cancer were eligible, with healthy volunteers recruited for comparison. Muscle biopsies of vastus lateralis were taken before and after 4-weeks home-based exercise (3x/wk). Mitochondrial oxidative phosphorylation (OXPHOS) was analysed using high-resolution respirometry (Oroboros Instruments). Results Eleven cancer patients and 6 healthy volunteers were recruited. Cancer patients had significantly lower maximal coupled (44.49pmol/(s*mg) vs. 74.81pmol/(s*mg), p=0.004) and uncoupled (55.82pmol/(s*mg) vs. 100.00pmol/(s*mg), p<0.001) OXPHOS capacity compared to healthy controls. Exercise training significantly increased maximal coupled (44.10pmol/(s*mg) vs. 54.64pmol/(s*mg), p=0.045) and uncoupled (56.20pmol/(s*mg) vs. 71.96pmol/(s*mg), p=0.001) OXPHOS capacity in this cancer cohort. Conclusion Declines in mitochondrial function appear to play a role in cancer cachexia. Short-term exercise training, in keeping with target timeframes for surgery (i.e., <31-days from decision-to-treat), appears to attenuate these declines. Further work is needed to determine if declines in mitochondrial function are causative or reactive, and to identify optimal exercise regimes to mitigate these declines.