RF is correct. The resistance of a semi-conductor will increase as temperature reduces. If you reduce the temperature far enough, it will stop working because there is no longer sufficient thermal energy to excite electrons into the conduction band.With all due respect, that is ridiculous. CPUs run just fine at -190 degrees C on liquid nitrogen cooling, as Tom's Hardware did here when they were able to overlock a P4 to 5.25 GHz at -196 degrees C: http://www.tomshardware.com/reviews/5-ghz-project,731.html
The important thing here is to be clear about where you are measuring the temperature. In the Tom's Hardware test, the -190°C was measured on the outside of the CPU case. The temperature inside will be much higher (by my rough calcs, for the junction temperature to be at about 30°C, the thermal conduction of the CPU would have to be in the order of about 2-7 W/m.K, which is probable, IMHO). If you poured liquid nitrogen onto the chip before turning it on, the junction temperature would get too cold and you would not be able to get it working. In the Tom's Hardware test, they had the CPU up and running before adding the cooling, IIRC.