Direct playback doesn't require much compute power. Just run Ubuntu server on the pi, spin up the containers, and get a fairly large HDD. An external HDD should be fine.

Any transcoding will be a issue tho. Like, if you use Jellyfin, and it wants to transcode your subtitles (even tho it shouldn't), you won't be able to stream anything 😂. I had this exact issue, and it was kinda pathetic that couldn't get subtitles to stop transcoding. It isn't transcoding now, but I also have a way more powerful server.

If you want to go more powerful. I'd recommend the build on Wolfgang's channel, with a N5105 NAS board, the N5105 is strong enough for 4k transcoding. It has 2x nvme, 6x sata, and up to 64gb of ram. Throw it in a decent case. I'd run Proxmox and Ubuntu server on Proxmox, this just makes it easy to backup your VM, in case something breaks and you want to rollback. At which point you can just throw HDDs in, or make them a ZFS pool, or a raid pool. Up to you.