Cloud computing has changed significantly in recent years because modern applications now require faster processing, lower delay, and better energy management. Traditional centralized cloud systems are often unable to support real-time services such as autonomous vehicles, smart healthcare systems, industrial automation, and large-scale IoT networks. As a result, researchers and industries are increasingly moving toward edge cloud architectures where data processing is distributed across edge devices and cloud servers. This paper reviews recent developments in edge-cloud collaborative computing, AI-driven orchestration, and sustainable cloud infrastructure. It discusses how technologies such as Deep Reinforcement Learning (DRL), Federated Learning (FL), and workload optimization techniques help improve resource management and reduce laten-cy. The paper also examines sustainable frameworks such as MAIZX and GEECO that focus on lowering carbon emissions and improving energy efficiency. Based on the reviewed studies, edge-cloud systems provide better response time, lower bandwidth consumption, and improved operational efficiency compared to traditional cloud only systems. However, challenges related to hardware heterogeneity, privacy, infrastructure cost, and AI explainability still remain important research issues.