Computers only work well in mild, stable-temperature environments. If the temperature swings wildly between daytime and evening (as it is apt to do in a garage), the computers will fail rather un-gracefully.

The underlying logic in this seems to relate to the ages-old, end-user quandry of "should we turn our PCs off at night?" I tried researching this at one point but what I found in the way of research wasn't very good. Certainly we would save energy and money by turning everything off, but the countervailing argument was that power cycles and hot/cold expansion and contraction would decrease component (mostly motherboard) longevity by repeatedly flexing traces. I was never very satisfied with the evidence one way or another.

High-end machine rooms and telco central offices, if they are not under-engineered WRT HVAC, run cool to the point that you need a sweater or jacket. However, the cooling usually overcompensates because in long rows of stacked/racked switches, even with proper installation, you can get some serious hot spots from exhaust convergence and keeping things chilly is a must to keep internals cool.

I wouldn't guess that it's the temperature in the garage that is such an issue but the ultimate resultant temperatures inside the case.. In your shoes, I think I'd see the garage as an irresistable opportunity. If temperature fluctuation was an issue, I'd probably monitor temperatures through the BIOS for a bit to get a baseline, then check to see what happens during garage temperature drops. One adjunct might be something like the thermal switch in the attached picture (from www.allelectronics.com in this case). If you could find the right normally-open switch, you could fix it so that your case fans only came on when the temperature inside the case reached a certain point. Conversely, if the temperature in the garage (and the computer) got chilly, the fan/s would shut off, allowing heat to build up inside the case until the thermostat was triggered again.