It’s more like, chips are tuned to operate in the same environment as we do.

Chips usually work better under cold conditions but may need higher voltages to compensate. That’s why extreme overclockers use liquid nitrogen. I had a phone I had to overvolt during winter for it to not crash when it goes too idle.

Heat affects conductivity and leakage currents especially in tiny circuits. More heat more conductivity, that’s why chips that run hot are sometimes susceptible to thermal runaway: they get so hot they conduct more which produces more heat until it fries. That’s also why chips being cooled better tend to reduce power consumption, or crash when they get too hot. Similarly, if they’re too cold, the given voltage may no longer be enough for it to switch fast enough and signals to make it all the way in time.

You can go really pretty cold if you give it extra juice. Just also have to account that the chip will produce a bit more heat, so depending on the cooling solution, the temperature differential between off and full power can be pretty big which would affect the ideal input voltages. It could need much more to start up but quickly require less as it heats up so it doesn’t immediately overheat. But with active temperature and performance monitoring like modern CPUs do, we could in theory expand the range a fair bit. Sometimes it’s easier to throw a heater at it.

Construction materials can also affect that, as metals shrink and expand in the cold and heat. That could cause the tiny wires to break especially if there’s big heat cycling.