There is a wide interest in biomarkers that capture the burden of detrimental factors as these accumulate with the passage of time, i.e., increasing age. Telomere length has received considerable attention as such a marker, because it is easily quantified and it may aid in disentangling the etiology of dementia or serve as predictive marker. We determined the association of telomere length with risk of Alzheimer's disease and all-cause dementia in a population-based setting. Within the Rotterdam Study, we performed quantitative PCR to measure mean leukocyte telomere length in blood. We determined the association of telomere length with risk of Alzheimer's disease until 2016, using Cox regression models. Of 1,961 participants (mean age 71.4±9.3 years, 57.1% women) with a median follow-up of 8.3 years, 237 individuals were diagnosed with Alzheimer's disease. We found a U-shaped association between telomere length and risk of Alzheimer's disease: compared to the middle tertile the adjusted hazard ratio was 1.59 (95% confidence interval (CI), 1.13-2.23) for the lowest tertile and 1.47 (1.03-2.10) for the highest tertile. Results were similarly U-shaped but slightly attenuated for all-cause dementia. In conclusion, shorter and longer telomere length are both associated with an increased risk of Alzheimer's disease in the general population.