Also I don't see the point of what TLS is supposed to solve here? If you and I (and everyone else) can legitimately get a certificate for 10.0.0.1, then what are you proving exactly over using a self-signed cert?

There would be no way of determining that I can connecting to my-organisation's 10.0.0.1 and not bad-org's 10.0.0.1.

Perhaps by providing some identifier in the URL?

ie. https://10.0.0.1(af81afa8394fd7aa)/index.htm

The identifier would be generated by the certificate authority upon your first request for a certificate, and every time you renew you get to keep the same one.

I see what you're getting at - but to me this sounds almost exactly like just using DNS, even if the (A/AAAA) record you want to use resolves to an un-routable address: https://letsencrypt.org/docs/challenge-types/#dns-01-challen... - you just create a DNS TXT record instead of them trying to access a server at the address for verification.

This is assuming NAT, with IPv6 you should be able to have globally unique IPs. (Not unique to IPv6 in theory, of course, but in practice almost no one these days is giving LAN devices public IPv4s).

A public CA won’t give you a cert for 10.0.0.1

Exactly - no one can prove they own it (on purpose because it's reserved for private network use, so no one can own it)

For ipv6 proof of ownership can easily be done with an outbound connection instead. And would work great for provisioning certs for internal only services.