Harnessing industry 4.0 technologies to reduce SMEs' carbon footprint: A machine learning-based assessment

Small and medium-sized enterprises (SMEs) significantly contribute to global greenhouse gas emissions, yet SMEs remain underrepresented in climate policy debates. Industry 4.0 (I4.0) technologies such as the Internet of Things (IoT), artificial intelligence (AI), and advanced analytics are increasingly recognized as critical enablers for reducing carbon emissions in manufacturing and service-based SMEs. Despite this potential, empirical evidence on their environmental effectiveness has remained limited, particularly in developing economies. In this study, the environmental impacts of adopting I4.0 for 120 SMEs operating in Türkiye were investigated, focusing on electricity, water, natural gas consumption and waste production. A mixed-methods approach was applied, combining structured surveys and operational records with descriptive and inferential statistics, scenario-based forecasting (2025-2030), and ensemble machine learning models (XGBoost, CatBoost, Random Forest). Carbon footprint equivalents were calculated using internationally recognized emission factors, and interpretability was enhanced through SHapley Additive Explanations (SHAP). The results were based on both empirical data and simulation-based estimates for missing resource consumption metrics. The results revealed statistically significant reductions in electricity (-2 %), natural gas (-8 %), and waste (-5 %) emissions post-I4.0 (p < 0.001). Scenario-based projections indicated cumulative emission reductions of 4.6 %, 9.2 %, and 17.3 % under low, medium, and high I4.0 adoption, respectively. Overall, the findings emphasized the role of data-driven digital transformation in enabling SMEs to achieve measurable environmental gains and underscored the need for targeted policy support, infrastructure investment, and capacity building to realize the sustainability potential of I4.0 technologies.