I like the idea and the end goal in theory. Based on my experience with how systemd was implemented I would be concerned that some use cases may not be thought

I like the idea and the end goal in theory. Based on my experience with how systemd was implemented I would be concerned that some use cases may not be thought through prior to this being mandatory e.g. keep it opt-in until kernel 7.x. I would hope that the first several iterations of this are opt-in so that people can discover blocking use cases they did not think of, such as PXE and iPXE booted systems, live kernel patching which is used heavily in some datacenters and especially in many VPS providers. The tooling around initrd image creation, enumeration and validation should be a first class priority in this effort in my opinion. This tooling should make it stupid-simple to enumerate the entire boot chain from UEFI to Initrd onward. i.e. walk through a simulation of the boot process and display missing dependencies prior to rebooting similar to the systemd dependency visualization and display cryptographic signatures through the entire chain. Another concern that is ping-ponging around in my head is the potential risk for abuse of this system. I could envision some Linux distributions combined with specific hardware models finding themselves one day locked into using Redhat/IBM approved boot loaders and images.