The Terraform AWS provider now supports predictive scaling policy, a machine learning-based scaling mechanism for Amazon EC2 Auto Scaling, allowing proactive capacity adjustments based on historical demand and usage patterns. This feature uses two metrics, Load Metric and Scaling Metric, to forecast demand and ensure enough instances are launched to meet predicted capacity requirements. Predictive scaling is designed to work in conjunction with dynamic scaling, setting a baseline capacity according to historic patterns while making upward adjustments for deviations from forecasted demand. To use this feature, users need Terraform v0.12 or greater and the latest version of the Terraform AWS provider, and can configure it by employing a new resource `aws_predictive_scaling_configuration` and modifying existing resources such as `aws_autoscaling_policy`.