This study develops nowcasting models for major components of Indonesian household consumption, namely food and beverage consumption, transportation and communication consumption, and housing and household equipment consumption. We utilize a rich set of monthly indicators (2013–2023) encompassing retail sales, credit, savings, inflation, and payment system data to predict current-quarter consumption in the absence of official data. Four machine learning algorithms (Elastic Net, Random Forest, XGBoost, and Support Vector Machine) are implemented and evaluated. The aim is to identify the best performing model and the key drivers for each consumption component. The results show that the Random Forest model yields the lowest nowcast error (Root Mean Square Error) for all three consumption components, outperforming the other methods. This superior performance is attributed to Random Forest’s ability to capture nonlinear relationships in the data. Using data available through March 2024, our nowcasts indicate stable year-on-year growth in food and beverage consumption (3.63%), transportation and communication (6.27%), and housing and household equipment (4.36%). These findings underscore the potential of machine learning-based nowcasting to provide timely insights for economic policy, enabling policymakers to respond more quickly to emerging consumption trends.