As an example of what I said earlier - take my 1988 Thunderbird Turbocoupe. When it came out, it was a marvel of modern tech (at least, for a North American car), one of the first with ABS braking and a whole host of computerized functions.
Today, if the ABS breaks, it's gone. There is no replacement, there's no way to replace or repair the computer that runs that. The ICs it uses are mostly unobtanium. The market is too small to make it worthwhile for an effort to reproduce or create an alternative. I wonder if that isn't the case too with vintage machines. Sure, the really popular ones might generate enough aggregate demand to make reproduction or such feasible, but when you get into the less popular machines.. like the Intertec Superbrains of the world, or even the Lisa... in 30 years are there going to be enough demand to make fabricating solutions worthwhile for individual machines? And at the end of the day, if you've changed the guts over to new programmable logic, say, is it even the same machine anymore?