Proceedings of the International scientific and practical conference ―Science, Technology and Culture in the Era of Globalization‖ (December 24-26, 2025) / Publisher website: www.naukainfo.com. – Geneva, Switzerland, 2026. – 234 p.

202 to reduce waste and lower operating costs [1]. Similar approaches are being actively implemented in industry to automate production processes, enabling significant energy savings through predictive analytics. In the transportation sector, CEAI supports route optimization, energy efficiency improvements, and predictive maintenance for electric vehicles and unmanned systems. CEAI plays an equally important role in agriculture, where it facilitates the implementation of resource-saving technologies such as precision irrigation, which reduces energy consumption and promotes environmental sustainability. The use of federated learning and hybrid Edge-Cloud architectures allows for timely anomaly detection, efficient energy management, and data confidentiality. Overall, the adoption of Collaborative Edge AI in various energy sectors and related industries creates the prerequisites for improving the reliability, efficiency, and sustainability of energy systems. The development of these technologies opens new horizons for the integration of renewable energy sources, optimization of resource use and the provision of secure, transparent, and flexible energy management in the face of growing modern challenges. Having considered the broad capabilities of Collaborative Edge AI for optimizing various aspects of energy systems, it's important to highlight the key challenges arising from its practical implementation in decentralized grids. As the number of edge devices, such as sensors, meters, and controllers, increases, the complexity of managing data flows and ensuring their timely processing also grows. At the same time, the variability of renewable energy sources, such as solar and wind generation, imposes stringent requirements on the adaptability and responsiveness of analytical algorithms. One of the most serious problems is system scalability, which is complicated by device heterogeneity in terms of different communication protocols, computational capabilities, and hardware architectures. To ensure effective interaction and coordination among such heterogeneous elements, it is necessary to develop flexible, modular solutions that enable devices to cluster into local groups based on shared