just exactly how would you propose to drive, say, 136 power FETs from an FPGA that operates on 1.2 volts and "tolerates" 3.3 volts?

That's one kind of task for which you'd want to be able to use an FPGA that lives in a couple of hundred pins. There are lots of FETs and FET drivers that like a 5-volt interface. It's only been in recent months that 1.5 volt compatible FETs have been available. I doubt that they're cheap, and I doubt they can handle the drive that's required to fire a solenoid or relay coil. The result is that one has to add buffers/level-shifters that then raise the system cost to far above what it would have been had the FPGA been omitted and simple SSI/MSI or MCU's been used.

This is not to say that they don't have their place, but I do think they've been overemphasized in educational institutions, while the material enabling students to make decisions as to when they're appropriate and when not, and, if not, what should be used in their place, has been entirely underemphasiszed. As a consequence, system cost and complexity goes up, and, as one would expect, system reliability goes down.

RE