Google Kubernetes Engine's Autopilot operational mode simplifies cluster management by handling provisioning and scaling. A recent upgrade enhances Autopilot's autoscaling with a dynamic, container-optimized compute platform. This allows rapid horizontal and vertical scaling based on Horizontal Pod Autoscaler and Vertical Pod Autoscaler configurations. Many customers, including Hotspring and Contextual AI, have benefited from Autopilot's operational ease and efficiency. In 2024, a significant portion of active GKE clusters were created in Autopilot mode. Now, Autopilot's benefits are available beyond dedicated Autopilot clusters to all qualified GKE clusters. This includes the container-optimized compute platform and simplified operations. These features are accessible through compute classes, with built-in options like 'autopilot' and 'autopilot-spot'. Users can even set Autopilot as the default for a namespace, optimizing bin-packing and resource utilization. Specialized compute classes allow running AI workloads on GPUs and TPUs with Autopilot's managed node properties. A new auto-provisioning mode for compute classes enables gradual adoption without impacting existing workloads. Google Cloud aims to provide these simplified and efficient operational capabilities to all GKE customers.