Corrosion is unlikely to cause any problem with the heat exchanger unless it is really severe, as there are no moving parts in it.

> nothing stops a circa-1994 Windows NT machine from continuing to work

My older machines of that era have all failed due (most likely) to failed capacitors. The disk drives fail most likely due to lubrication issues. The semiconductors fail because all semiconductors fail due to the migration of the doping that occurs when they get warm.

None of my 80's machines will power up anymore. I'm not too surprised, they were only built to last a couple years, and it only takes one failure in those enormously complex machines to do the whole thing in.

I try to keep my XP machines running because those were the last to support DOS programs and I still support Digital Mars C++ generating DOS programs.