Physical exercise is well known for its benefits on brain health. However, the mechanisms through which these benefits occur remain discussed, especially in the context of cognitive conditions such as Alzheimer’s disease. The present short review summarizes the findings of interventional studies that examined the effects of exercise training on the specific and non-specific biomarkers of Alzheimer’s disease. Controlled exercise intervention studies published in the English language were selected if they assessed the effects of a physical exercise intervention of at least 2 weeks in middle-aged or older adults on one of the following biomarkers measured either in the brain, the cerebrospinal fluid or the blood: beta-amyloid, tau, neurofilament light chain, and glial fibrillary acidic protein. Overall, there was no strong evidence of significant effects of exercise interventions on any of the selected biomarkers. However, in specific populations, such as women with obesity, pre-diabetes, or depression, favorable changes in blood beta-amyloid concentrations were reported. Further benefits on cerebrospinal fluid beta-amyloid were also demonstrated in APOE-ɛ4 allele carriers with Alzheimer’s disease. In conclusion, the current evidence suggests that physical exercise does not modulate the pathophysiology of Alzheimer’s disease in the overall population of middle-aged and older adults. Nonetheless, some specific populations, such as women with metabolic disorders and Alzheimer’s disease patients with APOE-ɛ4 genotype, seem to be favorably affected. Further studies, including long follow-ups, large sample sizes, and concomitantly assessing the effects of other factors such as sedentary behavior and diet, are required to bring further evidence to the field.