It also seems to be the case that CMSIS/DAP (the USB-level protocol for talking to SWD debug features) and/or debuggers are relatively chatty and inefficiency across USB, meaning that people see a noticeable improvement going from a debugger that only supports "full speed" USB (like the 32u4) to one that supports "high-speed" USB (like the UC32)
The Xplain series of development boards use an ATMega32U4 as a programmer/debugger- they call it the mEDGB.
The Xplained Pro boards have the UC32 EDBG chip.
The Xplained Mini boards have the 32u4 mEDBG.
The Curiosity Nano boards have a new SAMD "nEDBG" chip.
I'm sure if you wanted to get enterprising you could design your own board with an mEDGB onboard.
Sure. And it would probably cost about twice as much as a board without a debugger.
Those $10 development boards sold by chip vendors are nice, but they're not a very accurate indicator of how much it would cost to build a similar board on your own. Those boards are "loss-leaders"; the next best thing to a free sample; designed to attract devotees, and partially paid for from the "sales&marketing" budget...
I do wish that the current incarnation of Arduino would show more signs of having their manufacturing arm "in good shape." 