Spyder channels do not have the same accuracy specifications as the main MTUs, and they 'bottom out' (reading zero) at non-zero current flow/power flux levels which are dependent on the rating of the CT. I think that the TED folks assume that you want maximum accuracy on the main power and (where relevant) PV solar feeds but will tolerate the tradeoff between price and precision for branch feeds. Also, many people won't want the Spyder functionality.
Additional point added a few hours later: The Spyder hardware "piggybacks" on the main MTU hardware, which powers it and also does the PLC encoding and transmission of the data for it, so it is actually a pretty simple piece of kit. An MTU-less version would have to include hardware to replace these functionalities, almost certainly making it nearly as expensive to produce as the current full version.