I would rather be worried about something else than mere ageism.
Systems from the 1980s and earlier that we tend to consider definitely ‘retro’ are not merely old and obsolete. They are also very deeply publicly understood, down to assembly or even hardware level; not by random laymen obviously, but there is a body of publicly available knowledge that captures this intimate understanding, readily accessible to anyone who wants it and knows where to look, often for free. Of course, much of this is owed to the relative openness and simplicity of computing in earlier times: the number of possible configurations was relatively low, designs were often publicly documented (with manuals sometimes including source code, description of individual hardware registers, and even circuit schematics), and whatever wasn’t documented was easily reverse-engineered. It probably also helps that nobody is there to enforce NDAs from that time period.
As we move towards the modern era of computing, achieving this level of understanding becomes harder and harder. In the noughties, hardware manufacturers didn’t give you schematics and register listings, they gave you an opaque binary driver for operating systems they bothered to support, and if you happened to be running BeOS, tough luck. And operating systems became too large to feasibly analyse in their entirety and understand at a level that you might with Commodore 64 KERNAL.
I think it makes much sense to consider achieving this kind of deep understanding of a system to be a major aspect of retrocomputing, especially when it’s viewed as one of the facets of the DIY culture at large. After all, if your system is obsolete and therefore you can’t count on getting support from the manufacturer, you will need to service the system yourself. And to do that, you need to be able to understand how it works, not merely apply black-box solutions obtained from somewhere else.
Windows 9x can probably be counted among ‘pseudo-open’ systems thanks to extensive reverse engineering work by a number of people, but including even Windows 2000 here is considerably more dubious (unless you count leaked source code as ‘public knowledge’…). And I fear Windows XP slips away from that category even further. (This is something I also tried to allude to when answering whether KaiOS should be accepted as on-topic.)
So rather than trying to ascertain whether XP ‘has retro nature’ in itself, which is going to be mostly subjective, I’d rather ask what kind of questions about Windows XP we would be getting and whether we want them here. What I’d expect is that we’d see questions that approach the OS at a much more shallow, consumer level (much like How to make an audio CDROM for a(n emulated) Playstation?, which directs people to string together black box tools without understanding them), or that we’d get questions that are practically unanswerable without pretty laborious original research. I’m not sure I’d like either much, though the latter would at least seem more fun.