Of course in the limited cases where the application requires the heat to be transferred to the destination, you wouldn't use this method. Those pipes are likely heavily insulated.

But most cases, moving hot waste, the loss of heat is a side effect (the waste material just happens to be hot), and may in fact be an undesirable side-effect that squanders energy. Now, they don't need to squander that energy.

Easy example is a chimney - they're only trying to get rid of the exhaust gasses we don't want inside, and DGAF if it is cold or hot. In fact, heat exchangers are often installed to attempt to capture some of the lost heat and minimize losses.

Same for waste liquids going for treatment, or actual product exiting a hot process and going into a cold/ambient temp process -- they've got to dump the heat, far better to grab some energy on the way out.

> Easy example is a chimney - they're only trying to get rid of the exhaust gasses we don't want inside, and DGAF if it is cold or hot. In fact, heat exchangers are often installed to attempt to capture some of the lost heat and minimize losses.

Good example, except that you need a certain amount of heat to carry away the water (steam). I believe the rule of thumb for old-style furnaces is 80% efficiency with the 20% carrying away the water vapor.

Yup, and there's the same issue with using these on internal combustion automobile or truck exhausts - the density change of increased cooling could might change the back pressure/suction at the exhaust port and reduce the engine efficiency. It would surely be a consideration on highly tuned race motors, but I have no idea if the magnitude of the effect would be even noticeable on street cars & trucks. It is certainly fairly small, as you can drive through a puddle that soaks the exhaust and billows out clouds of steam, sucking out far more heat vs this thermoelectric system, and the engine doesn't even notice.

So, sure if you did it naively, without checking & compensating for secondary effects, you might have a net problem. But I'd be highly confidant that a bit of R&D and engineering design changes would easily take care of it.