Think of the that model-C hub as a splitter for one DisplayPort feed. Your Mac sends a single DP Alt Mode signal over the that model-C connection. Some hubs can convert that to two HDMI sockets but they are still sharing the same DP bandwidth. With DP 1.2 level signaling you have roughly enough for one 4K60 stream. The leftover becomes 4K30 or 1080p60 on the second port. Newer hubs that rely on MST to pack two 4K60 streams expect the computer to support MST for separate displays. macOS does not on external monitors. Not something macOS will magically enable with a firmware update. So for two 4K60 independent desktops from a base M2 you need a second display engine that does not come from native GPU output. That is what the that model graphics approaches provide. They present themselves as additional displays over that model and handle the HDMI conversion in the dock, which is why they need drivers. Thunderbolt without that trick will not bypass the one display limit on your model.
