Reading this it seems that Akami's numbers indicate if anything the available streaming speed of the connections.
The Ookla numbers indicate the speed of the connection under test conditions looking at throughput with a low load from other users on the connection.
So they are measuring different speeds and therefore are useful for different things.
If you want to know the linespeed averages the ookla numbers are an approximation. If you want to know throughput per application then Akami's numbers are the relevant ones.
So yes quite reasonable to doubt Akami's numbers for the linespeed (and therefore arguably the FTTP deployment) vs acceptable throughput (and therefore maybe the argument for minimum requirements)
This is not necessarily a contradictory position.