This is really not a good idea.
The economics of producing MCUs is dominated by two things (silicon area and pin count) and increasing either, increases the cost of the chip. This maybe doesn't make a lot of difference to a buyer of chips who buys a few hundred, but if you are buying millions, then it makes a significant difference. Also, debug is used only during the development of the end product - when shipping the product you want to switch that off - so do you want to bear the cost of USB debug in every device?
Historically, this (the debug problem) was solved by JTAG. When ARM developed their MCU's (Cortex-M), they cost reduced the debug by introducing SWD (smaller silicon area, fewer external pins, when compared with JTAG). SWD has a very simple interface that requires no on-chip drivers and takes very few pins. This interface is standardised and cheap, meaning that there are lots of cheap debug probes. MCU producers virtually give these away (LPC-Link2 is about $20) as it helps their MCU sales, but they are not keen on you using the same device for a different MCU vendor - through software. ARM addressed this by releasing CMSIS-DAP (which is available for most of the vendor-specific debug probes) and most debuggers/IDE now support this (meaning, for example, you can actually use LPC-Link2 with ST, or any other vendors debugger and silicon).
Note that if you buy, for example, J-Link for ARM, then you can use that on pretty much any ARM-based chip, regardless of the vendor. The reason J-Link s expensive is because it is expensive to produce and maintain, if you are not subsidised by selling MCU's. In my experience, J-Link is faster and much more reliable than the 'cheap' vendor-specific debug probes.
p.s. I'm nothing to do with NXP or any other chip maker, or debugger, just a (rather old, and long time) MCU developer.
