I'm happy to get 100 kbits/talker in that scenario with todays FHSS technology, 500 meters/hop, 1 watt power, 200 kHz channel width and 20 MHz of spectrum in the ISM band.
Of course, the advantage of Mesh Networks is they require little in the way of infrastructure, and scale up into the millions of nodes, with 50 million active node networks practical. But high data rates are not one of the properties of such networks.
a new circuit is established from O to B,
O tells A to tell the handset to move over to B-owned network infrastructure,
the handset roams over to B,
O drops the circuit to A.
This is a special case of Inter-MSC Handover, and you can read the spec at https://archive.org/stream/etsi_ts_123_009_v03.03.00/ts_1230...
The relevant section is Sec. 7.3.2, "Description of the subsequent handover procedure ii): MSC-B to MSC-B'".
(They call your "O" by "A", your "A" by "B", and your "C" by "B'" (B prime). And an MSC is a Mobile Switching Center that usually handles about half a city's worth of towers.)
Or the instructor I had taught us outdated information.
The cell phone is answered. So, it goes from switch O (the caller) to switch A (where your cell phone is associated with). Eventually, you'll move out of range of towers for switch A, and come into range of towers for switch B. But because the phone network is circuit switched, you can't just create a circuit from O to B. No, what happens is that A forms a connection to B, so now the call is going O -> A -> B. Talk long enough, and eventually, your call may end up going O -> A -> B -> C -> D -> E -> F.
So while it's true that your call doesn't go through multiple towers, it does go through multiple switches (as long as you are talking on the cell phone, and the cell phone is moving).