The point is the efficiency of the money spent on them for the open source ecosystem
Hence my question about SUSE and Canonical. I have exactly zero context for being able to determine that these expenses are excessive. They very well might, but "this number is bigger than the other one" without any industry context whatsoever just doesn't strike me as a meaningful argument.
That being said, if one's primary goal is to support open source development, the best way to spend one's money is obviously to donate to software projects directly. If one needs server support AND wants to spend money in a way that does most for development, the question still stands whether any direct competitors do any better.