To be fair to Microchip, their development tool chain is centered around a hobbled-down GCC port.
I don't mind paying well for good tools. (Granted, as I am a hobbyist, I am using free/open source tools with microcontrollers; but when I did run an IT company, I did unhesitatingly plonk down thousands of bucks at a time for software licenses for good tools, not to mention proper hardware, even though I was in my twenties and early thirties at the time. I knew even then, from experience, that it is worth the investment.)
The issue I've seen is that most of the professional software development tools are really not that good. The compilers do tend to be pretty good (and GCC in particular isn't excellent, just good enough), but the hardware abstraction and other libraries seem hell bent on keeping you within that specific ecosystem and never porting your code anywhere else. This is because that is in their interests to do so. To keep me buying new licenses, they need to keep pushing out New and Improved! versions, and keep maintenance costs of older versions as low as possible. Every bug they fix or not, is internally a value-cost proposition: how will this affect the profits from the software side?
On the positive side, these companies do not seem to be as short-sighted as pure software companies are. I guess that is because of the hardware guarantees of availability, and in general, longer-term projects than in the pure software world.