It’s a sound argument. Microchips will die, as will the circuit board they’re attached to. Even with lightly stressed electronics, you’ll see an ~80-90% death rate by about 40-50 years. Sure, there’s lots of 1970s quartz watches still around, but those are the survivors. It’s not like there’s a hard expiration date, it’s just a matter of statistics, so some will keep on going, just like there are 100 year old humans out there.

The difference for mechanical watches is that mechanical parts can be remade. Pivots can be repolished, a whole new balance arbor can be made by someone with a lathe, cracked jewel bearings can be pressed out and the holes refitted with new ones, and you can even oversize them a bit if needed to get a secure fit, and so on. Even a custom.hairspring or mainspring can be made from a coil of suitable spring material. Basically, all components can be made or remanufactured by a single skilled person, so it becomes a question of whether it’s worth repairing, not a question of can it be repaired.

Quartz can, to a degree, be repaired similarly, coild can.be replaced, mechanical parts fixed, a new circuit board, if stock is available, installed, but the similarity ends with the microchip. Even with advances in semiconductor manufacturing, we’re not at the point where someone can reverse engineer and duplicate a simple microchip using a workbench full of tools. It takes a factory with tens of millions to billions of dollars worth of equipment. We’re probably close to the point where someone could mount a chipscale packaged small microcontroller on a small custom-made interposer and program it to act like the original chip, but even that would be more expensive and difficult to do than making a custom balance arbor.

I’ve worked 30 years in the semiconductor and photonics world, developing devices, processes, and equipment used to make chips, supervising the torture testing and lifetime/reliability of chips, and so on. Chips all die at some point. They start un-making themselves the instant you power them on. Non-integrated electronics, like the single transistor circuit in an old Accutron, can potentially be repaired with alternate components, since there’s always going to be an assortment of individual transistors, diodes, capacitors, and whatnot made. It might be difficult and require experimentation to get a sufficient match, and there’s no clear replacement for things like the coils, but, overall, they’re much less stressed due to relatively large circuits handling relatively small amounts of power, so they’ll live longer to begin with, and they’re made of simple enough parts that are human replaceable, so they might be with us longer than early quartz of the same era. But, still, electronics die, and they die in non-repairable ways. Integrated circuits, like those in quartz modules, die faster due to being smaller structures handling similar power levels.

Mechanical systems do more or less the same thing, un-making themselves in the act of operating. But, their scale is much larger relative to the power they handle, so they live longer. But, they do accumulate wear, and will fail in time, plus the can be damaged in specific ways, like water, or a good hard knock. However, precision shaping of small mechanical components can be done by a single person with benchtop tools, making them repairable, and a microchip can’t, at least not yet.

That’s the wildcard here. Is it possible to imagine a benchtop tool set that would allow simple microchips to be made by a solo engineer? Absolutely. But what’s the market demand? Is it worth the billions to develop such hardware to sell a few $50k tools to weird hobbyists? Probably not. But, it’s possible that a weird hobbyist builds such capability on their own. There’s a lot of talented hardware hackers out there, like Sam Zeloof, who, a few years ago, managed to do just what I’m describing, assembling his own basic chip making capability in his garage, ultimately producing a handful of DIY versions on Intel’s 4004 CPU from the 1970s. That’s a hell of a feat. But, drivers/counters for quartz watches are a different story. The 4004 was well documented and was made in large quantities, so he had a lot of reference material to work from. This may not be true for many quartz timer controllers.

In any case, it’s speculatively possible that, in the future, it could be easy and commonplace to remake old chips using readily available technology, but I wouldn’t count on it, and without that, vintage quartz movements are doomed to fade away and become inert antiques at some point. Of course, this is true of mechanical watches, too. Many become not worth fixing at some point, or are so difficult to fix that you don’t use them for fear of something terrible going wrong. I have a couple of old, late 1700s, verge fusee watches that sort of run. If you wind them, they’ll tick for an hour or two. But I won’t bother fixing them. They’re not exceptional examples of their type, and would be expensive and difficult to repair. The pivots are worn, as are the simple metal bearings, the mainsprings are old and have “set”, no longer producing much torque, and so on. Plus, my wardrobe is entirely lacking in embroidered waistcoats and powdered wigs, so I have nothing to wear with one. They’re better as curios on the bookshelf than they are as functional watches.

So, quartz has, for a number of reasons, a shorter fuse to its existence than most mechanical watches, but every technology, whether silicon or metal based, ends up no longer functional someday. As with us, so with our creations: ashes to ashes, dust to dust, sand to sand, rust to rust.