I have mixed feelings about it. It does take some money to run the site, but there are some
very ambitious folks there who want to make it more of a big-time NGO, so much of the fundraising is not really aimed at servers+code, but at some of the big-dream initiatives, like fixing education in Africa or something. There are also increasing numbers of "overhead" employees: full-time donor-relations and PR/marketing/outreach folks (not just one or two of them, but more like 10-15). They have picked up technical hiring, though; for a while,
all the new hires were of the other sort.
I used to be really involved in Wikimedia meta-stuff from 2003-2006 or so, until more of that started happening. I personally favor a grassroots-community plus minimal legal infrastruture model, like Debian has with its "parent" foundation Software in the Public Interest, which basically pays the bills and otherwise mostly stays out of the way. Other people, especially those with previous experience in the "nonprofit sector", have in mind a much more professionalized nonprofit organization, along the lines of the Sierra Club or one of the big NGOs. I'm skeptical of those organizations, especially when it comes to community-driven initiatives, though I think they do sometimes do good work. But imagine a Debian where the Release Manager position was actually a full-time, professional position, hired by SPI. There are some pros to that (a full-time employee being paid to do the job can devote more time), but I think it runs a lot of risks as well.
Some of one's views on that might also depends on what you think the scope of the Wikimedia Foundation should be. I tend to favor a more limited scope, where they focus on shepherding the production of content (by volunteers), and then making it available for free under an open-content license. Other people favor a more comprehensive mission of being a content producer+publisher+distributor. There is much more that can be done with the content once it exists, in terms of repackaging, distributing it to various parts of the world, remixing it into useful subsets, producing new interfaces to it, etc.
But I'd rather see third parties do that: part of the whole point of making something open content is that anyone can do those things, not only the Wikimedia Foundation, and you don't even need its permission to do them (though friendly coordination is always good). I think the whole ecosystem would be healthier if there were more interesting things being done with the Wikipedia content by organizations and people other than the Wikimedia Foundation. Instead of doing those things itself, I'd rather it focused on figuring out why not that many third parties are doing interesting things with the content (one potential point of improvement: what kinds of database dumps Wikimedia provides, and how regularly it provides them).