A targeted perspective synthesizes epidemiological and mechanistic evidence examining how caffeine consumption during midlife may influence dementia risk later in life. The authors highlight potential roles for dose, age, CYP1A2 genotype, adenosine signaling and cultural drinking practices, while stressing that current observational evidence cannot establish causality.