But that's not how it's supposed to work, according to Ledger. They state that the seed phrase undergoes encryption and is divided into three shreds. These shreds are then directly sent to the three custodians from the Ledger device itself. When a recovery is requested, these encrypted parts are sent back to the new or old hardware device and decrypted back in the recovery seed. Nowhere does it mention that the shreds must pass through any Ledger server for encryption or decryption during recovery. Additionally, the process you described would imply that Ledger stores all private encryption keys from every device they have ever produced on their servers, which would create a single point of failure. It wouldn't make sense to keep such a system in place, and the entire process of splitting the recovery seed into shreds and distributing them to three different custodians wouldn't make sense in that case.
I just described an example of how this can be implemented. All Ledger nano X wallets have unique bluetooth identifiers, so what's stopping them from adding something like that to security chips? It would also be a great stupidity on their part to transfer the seeds in an unprotected form. When they launch this service will be more clear, now one can only guess.