The best breakdown I have seen on this matter was this op-ed by a long-time and quite prominent Wikipedian on Wikipedia Foundation (WMF) spending patterns.

Here's an excerpt:

"The modern Wikipedia has 11-12 times as many page views than it had in 2005,[23] but the WMF is spending 33 times as much to serve up these pages to the readers.[24] This seems reasonable given that they have improved reliability, redundancy and backups. More concerning is the fact that since 2005 the WMF has hired hundreds of extra employees and is now spending 1,250 times as much overall,[25] which seems rather excessive considering that the actual amount of work they have to do is pretty much the same. WMF's spending has gone up by 85% over the past three years.[26]"

The point that this Guy is making is that the rate of spending increase is concerningly high, and that it isn't showing signs of levelling off (at least, as far as the data in the article, which goes up to 2025, shows). He's concerned that if spending continues to increase at such a high rate, then it may overtake donations, causing WMF to have to dip into their endowment.

Now, that wouldn't be a huge problem by itself, but what Macon fears is that, should this happen, WMF wouldn't adapt appropriately and they'd just end up in a spending death loop that would lead to the loss of Wikipedia as we know it (and imperil all the other cool stuff they do, like those you list).

I don't have much personal experience editing Wikipedia myself, but I have plenty of friends who do, and what I've heard from them leads me to take Guy Macon's warning seriously. Apparently Wikipedia is quite bad on the software development side of things, as well as spending transparency.

"If [WMF] had [followed standard software engineering principles], we would be seeing things like requirements documents and schedules with measurable milestones. This failure is almost certainly a systemic problem directly caused by top management, not by the developers doing the actual work.

I remember the development and deployment of the Visual Editor causing particular hubbub amongst my nerdy Wikipedia friends, and Macon cites this as an example:

"Take as an example the way Visual Editor was developed. There are many pages of documentation on the WMF servers and mailing lists, but no evidence that any developer had any serious discussions with the actual editors of Wikipedia who would be using the software. Instead, the role of "customer" was played by paid WMF staffers who thought that they knew what Wikipedia editors need better than the editors themselves do. Then they threw the result over the wall, and the community of Wikipedia editors largely rejected it."

When I dug into this further though, the biggest concern for me was the lack of spending transparency. This article by Guy Macon blew up because a lot of people shared his concerns, and those people naturally wanted a clearer accounting on how money was being spent. The fact that that has continually resulted in little, if any improvement or fulfillment of promises made is a bad look for an organisation that claims to pride itself on radical transparency.

Because this essay is hosted on Wikipedia, there's a heckton of other discussions on the talk page for that essay too, if you want to check it out.