In sport, the awakening responses of cortisol (CAR) and testosterone (TAR) have been used as evaluative tools. Research findings are, however, inconsistent and the mechanisms involved are unclear. This study investigated the CAR and TAR in male athletes across light and heavy training weeks, focusing on skeletal muscle damage as a mediating factor. Twenty elite male judokas were assessed across consecutive weeks of light and heavy training (i.e., 6 days, 9–10 weekly sessions). Plasma cortisol and testosterone concentrations were measured post-awakening (+3, +30, +60 mins), along with creatine kinase (CK) at +3 mins. The CAR and TAR were indexed by baseline-corrected change scores (Δb30, Δb60) and area under the curve (AUCb30, AUCb60). The early-morning surge in plasma cortisol concentration (CARΔb30, CCARΔb60, CARAUCb30, CARAUCb60) was significantly larger after light versus heavy training with effect-size differences of 2.14–2.64. The post-awakening decline in plasma testosterone (TARΔb30, TARAUCb30, TARAUCb60) was found to be significantly less pronounced, whilst CK level was elevated, after heavy than light training with effect-size differences of 0.95–1.04 and 4.70, respectively. Causal mediation analysis confirmed that CK mediated, in part, the training effect on the CAR, but not TAR, measures. In summary, male judokas, upon rising after a light training block, displayed a rising CAR (36%, 22%) and declining TAR (-11%, -15%) at +30 and +60 mins. A high-intensity training block suppressed the CAR (21%, 8%) and attenuated the TAR (-7%, -13%) with accompanying muscle damage offering one mechanism to partly explain the CAR differences.