Energy-efficient manufacturing has become a central focus in the industry due to sustainability concerns and supply volatility. Due to their energy intensity and wide use, injection molding machines (IMMs) account for a significant share of global manufacturing energy consumption. Existing models for IMM energy estimation primarily focus on individual machines, making model scalability difficult in cases where manufacturers use diverse machine types. This paper presents a data-driven approach to model the energy consumption of IMM modules (i.e. a collection of IMMs in a production facility), accounting for the collective energy consumption of auxiliary systems such as heating, cooling, drying, and ventilation systems. Using production data from a large manufacturer, we identify the minimum set of production variables (features) needed for accurate energy consumption modeling. Key features include the number of machines, machine size, machine type, and shot weight. These features were used to train random forest-based machine learning models to estimate the energy consumption of 3 separate IMM modules. Results show average \(R^2\) values of 0.95, 0.94, and \(-0.63\) for models trained on 80%, 20%, and 0% module-specific data, respectively. Our findings suggest that the significance of production variables remains consistent across IMM module types. However, module-specific training data is necessary for accurately estimating module energy consumption.