OBJECTIVES: The economic burden of long term care for people with dementia is significant and substantially covered by state Medicaid programs. While no therapies exist to substantially alter disease progression, supportive services for caregivers have been shown to improve caregiver capabilities and well-being and delay patients’ institutionalization. We estimated the potential cost savings from offering the New York University Caregiver Intervention (NYUCI) to eligible Medicaid enrollees from a state payer perspective. METHODS: Data from the Minnesota Department of Human Services and the literature informed a population-based Markov model to predict and compare costs, including intervention costs, over 15 years with and without implementation of the NYUCI for family caregivers of community-based Medicaid eligibles with dementia. We incorporated original NYUCI randomized controlled trial results as appropriate, adjusted to Minnesota implementation. RESULTS: Approximately 5-6% more eligibles with dementia would remain in the community from year 3 (2014) on and 17% fewer would die in nursing homes over 15 years if their caregivers received the NYUCI. The state could save $40.4 million (2011 dollars, discounted at 3%) if all eligibles/caregivers participated in the NYUCI. Savings are expected within 5 years of program implementation. Best-and worst-case scenario estimates were $178.9 million and -$7.3 million, respectively, driven largely by the variation in the assumed NYUCI effect. CONCLUSIONS: State payers can use enhanced caregiver support to moderate the growing tax burden of dementia, even without a breakthrough in pharmacologic treatment of the disease. But because it takes time to realize savings from a program like the NYUCI, the frequently short time horizon of legislative fiscal notes presents a challenge to adoption. In deciding on program value it may be important to distinguish cost-effectiveness from return on investment and take long-term impact into consideration.