BACKGROUND:Frailty and social isolation are independently associated with dementia risk. This study examined whether frailty predicts dementia differently for older adults living alone versus with others. METHODS:Participants were 967 community-dwelling, dementia-free older adults from the Sydney Memory and Ageing Study (2005-2020). Frailty was assessed using an adapted Fried Frailty Phenotype (robust, pre-frail, frail). Dementia was diagnosed via biennial neuropsychological testing and clinical consensus. Cox regressions analysed associations between baseline frailty, living arrangements (alone vs. with others), and incident dementia over 12 years. RESULT:Frail individuals had 1.74 times higher dementia risk than robust individuals (HR = 1.74, 95% CI: 1.16-2.60, p = .007). Living alone was not independently associated with dementia (HR = 1.03, p = .825). The interaction between frailty and living arrangements was non-significant (HR = 1.23, p = 0.600). Exploratory stratified analyses showed frailty predicted dementia among those living alone (HR = 1.92, 95% CI: 1.06-3.51, p = .033) but not those living with others (HR = 1.42, p = 0.224). CONCLUSION:Frailty is a significant predictor of dementia risk in older adults. While the formal interaction test was non-significant, exploratory findings suggest frail individuals living alone may be particularly vulnerable, warranting further investigation and potential targeting for combined physical and social interventions, pending replication.