GPUs dump a redonkulous amount of heat in a very small space. I imagine a passive cooler for one might easily take up more room than the entire cubesat.
Another factor is lubrication (mentioned in passing by Doctor Jackson above). Lubricants that work in an atmosphere do not always work in a vacuum, so NASA had to develop some new lubricants for machinery in space. Since the Tesla does not have these special lubricants, it would probably only run for a short time before something grinds to a halt.
ETA: chances are that the GPU wasn’t space-rated anyway. Orbit has a higher radiation background than on Earth, so you need processors that can handle that. otherwise you keep getting a lot of faults and other problems.
This was a fairly small GPU in an integrated package; on the order of 10 watts. That is low enough that a ~20 g heatsink with a dinky fan does the job in air. In a vacuum, it would overheat within a minute (the aluminum heatsink would initially soak up the energy via its thermal capacity, but not have any way of getting rid of it once soaked).
Some heatpipes coupled to the frame, maybe with some extra conduction paths to the solar panels probably would have been enough (though I never did the math). This group had enough other problems that it was academic, though.
Not really necessary. An ordinary tire might be inflated to 35 psi. In a vacuum, that’s like 50 psi. That might be higher than the tire’s rating but it’s not enough to burst it–the tire has plenty of safety margin to account for the usual bumps and other variation when driving.
Low Earth Orbit is surprisingly gentle. Our cubesat used an ordinary non-hardened microcontroller and survived nearly a month without a reset (might have gone longer but that was our only contact).
Of course, if you need something with high continuous reliability, you either pick something hardened or build in redundancy (SpaceX uses this approach). But cubesats don’t need this level of reliability and as long as you have a watchdog circuit that resets the thing if there’s a hang, there’s really no problem at all.
The thermals really are the hardest part. The radiation is not something that cubesat designers worry much about.
Outgassing from the electrolytes, plastics and other materials inside a standard lithium battery might be a problem in a vacuum or low pressure environment; this probably happens a certain amount even under normal conditions, but the casings will be designed to handle these things at STP.
nm