A faster transmission of a CAN FD frame's payload reduces its transmission time, i.e. the bus load decreases. On the other hand a longer data field could be transmitted in the same time, i.e. with only a single CAN FD frame.
If the data field is transmitted for example at a fivefold speed with CAN FD, the frame would have approximately the same length in time as a normal CAN frame if it contained five times more data. Therefore a CAN FD frame can transport five times the information of a CAN frame without raising the busload.
Introducing CAN FD would result in a number of advantages for the development process:
To make CAN FD work, new CAN FD controllers are needed. These however are downward compatible and are able to handle classical CAN. ECUs on a CAN bus can be replaced with CAN FD capable ones step by step. Certainly traditional CAN must be used on any bus as long as there is still at least one ECU with a conventional CAN controller.
So it appears as if the benefit of CAN FD can only be used after all ECUs on the bus have been upgraded with CAN FD controllers. But this impression is not quite correct. With only some ECUs being CAN FD capable they can communicate at higher speeds while the others have been put asleep via partial networking. This can be used for example for updating ECU software during service. Especially there higher data rates are advantageous because update processes often take hours.
Another advantage of CAN FD is that it does not involve a completely new technology. All CAN-related knowledge and experience acquired by engineers over years of working with CAN, is still valid and can be reused with CAN FD. This implies that the process of upgrading from CAN to CAN FD can be achieved with neither huge costs for education nor the loss of experience.