As collaborative robots (cobots) become increasingly common in industrial settings, managing their energy consumption is critical for sustainable manufacturing. This study addresses the problem of accurately predicting the power usage of a UR10e cobot during a 3D printing process, where the robot served as a heated platform. Using real-time telemetry data such as joint speeds, voltage, and current a regression model was developed within the KNIME Analytics Platform using the XGBoost Tree Ensemble Learner. Time-series features were engineered to capture the robot’s motion dynamics, enabling effective estimation of energy consumption from movement data. This study demonstrates the applicability of low-code machine learning tools for robotic energy forecasting, with potential for extension to diverse tasks within Industry 4.0/5.0 environments.