Alzheimer's disease (AD) is the most common form of neurodegenerative disorder with dementia. In its sporadic form, AD results from the combination of genetic factors with different epigenetic events. Among them, oxidative metabolic reactions and their by-products have been consistently implicated in AD pathogenesis and represent the biological basis for the 'oxidative stress hypothesis' of AD. Numerous studies demonstrate that different biomarkers of oxidative-stress-mediated events are elevated in the AD brain. Studies in animal models of the disease with antioxidants report significant improvements of their AD-like phenotype. Although epidemiologic studies show that dietary intake of antioxidants reduces the risk of AD, clinical trials with antioxidants show only a marginal positive or no effect. These conflicting results have created a wave of criticism towards the oxidative stress hypothesis of AD. Here, I review the available data and discuss the necessary paths for a fair reappraisal of the hypothesis.