I don't get why they did this, Ark is such a great resource. If there is any cost to keeping the data usable, surely it can't be that much worse for ultra-legacy products that it only makes sense to keep around "newer" (and I use that term very generously) products.
You know they've got at least 10x the server cost being wasted by old emails and stuff like that which could be more sensibly purged instead of customer reference documents.
Perhaps it's an anti-recycling effort against those who would otherwise keep running these old processors which once played a respectable key part in Intel's progress.
After digital files first became available for electronic components, there has never been any good excuse for them to go out-of-print after that.
If it isn't just some error
Turns out like every corporation, they have both sides.
Second, for compilers you have projects like LLVM that's like a framework for building compilers where those companies like Intel, AMD, Samsung, etc contribute.
LLVM shouldn't turn into compiler monoculture.
Even if they had backups[1], that doesn't necessarily mean they'll remain publicly accessible. I learned that the hard way many years ago when my first university Unix account web pages from the 1990s were made inaccessible, effectively permanently AFAICT, after the university IT published a robots.txt on that domain that caused (deliberately or not) Archive to hide them, and then later discontinued use of that domain altogether. I think Archive has since changed their policy regarding robots.txt, but the lesson remains the same.
Ever since then I try to remember to archive (e.g. wget -m) important third-party pages and artifacts (PDFs, etc) on my personal server, especially reference materials, memorable articles and blog posts, etc. I wish I had that foresight earlier. Though even before that happened I had already started hosting all my own stuff.
I still find Archive's Wayback Machine to be immensely valuable and useful. But if you find yourself going back to the same source over and over again, do yourself and everyone else a favor and archive it yourself, in addition to adding it to the Wayback Machine if it's not already there.
[1] The E8400 page is presently available, at least: https://web.archive.org/web/20230927192341/https://www.intel...
But if you're searching for now defunct pages it's hard unless you know the url already. eg if I wanted the page to the q6600 for example, how do I even get that?