The rapid expansion of solar energy infrastructure has necessitated advanced grid management solutions to ensure reliability, scalability, and efficiency. Conventional power grids struggle with fluctuating solar energy generation, leading to power imbalances, voltage instability, and increased downtime. Additionally, traditional forecasting methods lack the accuracy required for optimal energy distribution, while conventional fault detection systems exhibit significant response delays. To address these challenges, this paper proposes a Cloud-AI Smart Solar Grid Framework that leverages artificial intelligence (AI), cloud computing, and machine learning (ML) for real-time energy forecasting, grid stability optimization, and autonomous self- healing mechanisms. The proposed system integrates Long Short-Term Memory (LSTM) networks for solar energy forecasting, reducing mean absolute percentage error (MAPE) to 4.2%, a 60% improvement over traditional methods. An AI-driven demand- response optimization mechanism minimizes power deviation by 45.3%, significantly enhancing grid stability. Furthermore, the self-healing grid mechanism detects and mitigates faults with 98.7% accuracy, reducing grid downtime by 85.3% compared to traditional SCADA systems. The cloud-integrated architecture enables large-scale, decentralized power grid operations, maintaining low-latency processing (below 50 ms for 1000 + nodes) while ensuring resilience against dynamic load fluctuations. The results demonstrate that AI-driven solar grid intelligence and cloud-based optimization significantly enhance energy efficiency, operational reliability, and grid resilience, making it a viable solution for next-generation smart energy systems.