Kind of, Keda scales to zero by scaling replicas. Meaning when a request comes in it spawns a new Pod, which can have significant delays (image pulls, application start etc.). It also requires a central layer 7 proxy, which zeropod avoids. Disclaimer: I created zeropod