"Serverless Endpoints" aka "Scale to Zero"

### Describe your suggestion

I would like to know if serverless endpoints or scale-to-zero endpoints are something that the Azure team is working on. I think I don't need to get into why this feature would be beneficial to have and I believe AWS rolled out something similar earlier this year.

Is it on the roadmap? When is it expected to come?

EDIT: To clarify I'm not talking about auto shutoff for compute instances. I'm talking about something similar for inference endpoints.

Thanks,

### Additional details

_No response_