CloudAIBus is a new testbed for AI-driven cloud computing environments designed to improve resource allocation efficiency and sustainability. Traditional methods often over-provision CPU resources, leading to financial inefficiencies. Recent AI techniques, such as deep learning, can accurately forecast resource usage, enabling more efficient allocation. However, the dynamic scaling potential of AI models has not been thoroughly explored. To address this, CloudAIBus was developed, employing the DeepAR model for CPU usage forecasting to support cost-effective resource allocation decisions. The DeepAR model was implemented using Amazon SageMaker, a robust platform for scalable training. Performance was evaluated using Google Colab, with results showing better performance than baseline methods (LSTM and ARIMA-based resource management) in terms of MAE, MAPE, and MSE. The proposed approach reduced unused CPU percentage from 98.65 to 32.35% compared to the GWA-T-12 dataset, demonstrating its effectiveness in reducing over-provisioning. CloudAIBus aims to address economic and environmental challenges by providing a scalable and adaptable solution that enhances cloud efficiency and sustainability. It integrates advanced AI models to improve predictive accuracy and operational efficiency, aligning with global efforts to achieve sustainable IT practices. The system offers significant implications for the cloud computing industry, spanning financial, operational, and environmental aspects. CloudAIBus addresses the need for sophisticated resource management solutions in complex cloud environments, promising advancements in resource allocation and management. The main contributions include developing a new testbed for AI-based resource management, utilizing DeepAR for CPU forecasting, implementing DeepAR on Amazon SageMaker, evaluating performance using a dataset of 1750 VM traces, comparing performance with existing methods, and measuring CPU utilization efficiency.CloudAIBus is a new testbed for AI-driven cloud computing environments designed to improve resource allocation efficiency and sustainability. Traditional methods often over-provision CPU resources, leading to financial inefficiencies. Recent AI techniques, such as deep learning, can accurately forecast resource usage, enabling more efficient allocation. However, the dynamic scaling potential of AI models has not been thoroughly explored. To address this, CloudAIBus was developed, employing the DeepAR model for CPU usage forecasting to support cost-effective resource allocation decisions. The DeepAR model was implemented using Amazon SageMaker, a robust platform for scalable training. Performance was evaluated using Google Colab, with results showing better performance than baseline methods (LSTM and ARIMA-based resource management) in terms of MAE, MAPE, and MSE. The proposed approach reduced unused CPU percentage from 98.65 to 32.35% compared to the GWA-T-12 dataset, demonstrating its effectiveness in reducing over-provisioning. CloudAIBus aims to address economic and environmental challenges by providing a scalable and adaptable solution that enhances cloud efficiency and sustainability. It integrates advanced AI models to improve predictive accuracy and operational efficiency, aligning with global efforts to achieve sustainable IT practices. The system offers significant implications for the cloud computing industry, spanning financial, operational, and environmental aspects. CloudAIBus addresses the need for sophisticated resource management solutions in complex cloud environments, promising advancements in resource allocation and management. The main contributions include developing a new testbed for AI-based resource management, utilizing DeepAR for CPU forecasting, implementing DeepAR on Amazon SageMaker, evaluating performance using a dataset of 1750 VM traces, comparing performance with existing methods, and measuring CPU utilization efficiency.