Until it doesn't.
As someone who has worked in third-party repair and calibration for longer than I care to remember, I can tell you that this crap happens all the time, with all sorts of vendors/manufacturers. Manufacturers have little interest in what they are not currently manufacturing. And your so-called "extended warranty" is a thinly-veiled euphemism for insurance. They are betting that enough people will buy that insurance and not need it, so they don't lose money on the ones that do.
Some companies make legal threats over third-party repairs on equipment that is way past EOL (End Of Life). One might think, well, what skin is it off of their nose? They believe that corporations will fork out the cash for a newer model (or whatever) if they can shut down the repair of older, perfectly adequate equipment. Companies with this equipment sometimes don't have the resources to just up and buy the latest and greatest, and then spend the down-time on their production lines (again, or whatever), what with all the installation of the new stuff, the rewriting of software (consider multiple ATE [Automated Test Equipment] racks that would need to designed almost from the ground up to adopt newer standards, for instance), and the re-training of the employees for the new stuff. All when the equipment they currently own is perfectly fine for what they are doing. So, they go to a third party. In most cases, getting schematics can be a nightmare, and the same goes for parts. We often have to almost reverse-engineer stuff, just to figure out what the original engineering team was trying to do. Gah.
A related aside: I once maintained and old Sun Microsystems "pizza-box" server (a SPARCstation 10, circa 1992 IIRC), that didn't do...ANYTHING. I kept it up and alive so it could be pinged on the network. They needed it to be up and functional "just in case" they ever needed it (because you see, the FAA required that all equipment used for airworthy certification be available for X number of years after the original certification). So, this thing just sat there running, for years and years, until the pigeon-poop (thermal compound) between the CPUs and heatsinks dried out (and three of the four cooling fans died of old-age, LOL) and the spring-loaded nylon screws holding the heatsinks finally popped off due to heat. Of course, the CPUs fried immediately. And I had to fix it. This was about ten years ago. Now, guess what would have happened in this current era's wet-dream of complete control of manufactured goods?
Yes, it used to be IBM, then Microsoft. Now, it seems to be Apple, although Google is certainly giving them a run for their money.
There ya go.