That's easier said than done. The air protocol (the data format that is sent out on LoRA radio) for each have different fields, of different sizes, with different semantics. MeshCore did routing from day one but they have their own issues with a one byte repeater address. Meshtastic also had roles defined for routers, but it's not exactly usable in the way envisioned.

This is not to say that it couldn't be done, but it's likely to end up being an https://xkcd.com/927/ situation. More likely, the two would remain relevant for their respective strengths, with a possible higher-level routing application that can pass messages between the two. This is no different than the BlueSky/ActivityPub bridge between those two networks.

There is also the prospect that a non-LoRA protocol may appear, possibly one which is built atop 802.11ah HaLow, for example. The history of the rise of both Meshtastic and MeshCore required cheap, mass-produced, available LoRA chips. There's no reason the same could happen in future, especially when it's a formalized open-standard that takes the best from all existing protocols (eg MeshCore, Meshtastic, WiFi, BLE).

Remember that LoRA is, at the end of the day, a technology currently produced in silicon by a single company, something the LoRA Alliance readily admits in their FAQ. This is an exposure which an open-standard (eg IEEE standards, where the specification can be purchased for a ludicrously high price, and then there are no barriers to manufacturing custom silicon) does not have.