Why did the liquid seemingly disappear from some unopened aluminum cans? I'm sitting in a room next to some totally unopened cans of carbonated soft drinks (if it matters — the two affected cans are Coke Zero and Diet Coke). These cans were unboxed approximately two days ago and were full, but now are seemingly mostly empty!
There is no puncture in either can, and compressing them results in the dent being counteracted by pressure inside the can. Shaking and weighing the can (by hand; I don't have a scale anywhere near) reveals that a good majority (75%) of the can is now effectively empty.
The cans spent the whole time stored inside a normal room with a temperature of ±22ºC and automatic fluorescent lighting which is on for 16 hours a day.
Finally, there is no condensed drink on the outside wall of the can, and the cans themselves are externally room temperature, so I'm really puzzled about what process could do this.
I'm hesitant to open the cans because I'm not sure what the result of that could be, however if it's deemed here that there's little or no risk, I'll do that and update the question.
What possible reason could there be for these cans becoming empty?
 A: Count_to_10's explanation is good but it does not seem to fit with all of the circumstances which you have described.
It seems strange that the 2 cans were "full" after being stored 20 months beyond their sell-by date, and then "mostly empty" only 2 days later. The only thing which has happened in between is that they were "unboxed." It seems to me unlikely that unpacking could have triggered a leak - a leak is much more likely to be triggered by rough handling during delivery to or handling by the retailer.
The fact that the cans are still under pressure seems incompatible with a leak. Gas (with smaller molecules) will leak more easily than sugary liquid. The internal pressure will enhance the leak, and the can will gradually lose pressure. However, you say that the cans can be "compressed" and "dented" so I think the internal pressure might in fact be quite low.
It also seems strange that there is no sign of any sticky liquid outside the can. The sugar and coloring will not evaporate, so there should be a sticky brown ring around the bottom of the can, as there was here and here. (There is a very detailed post at the end of the latter thread, including results of a home experiment and some research on the corrosion of aluminium.)
Did you check the packaging for signs (rings) of leaked-out cola?
A faulty seam at the top of the can could explain a leak of gas and water vapour and the absence of a brown residue. But it cannot explain how so much liquid has evaporated through a microscopic hole in only 2 days. So much liquid would not evaporate in 2 days at 22$^{\circ}$C even if the can were open.
I think more likely explanations are that either (a) the cans which leaked did so slowly over months or years, and were not the ones you tested and found to be "full" when you opened the box 2 days ago, or (b) there was a filling fault at the factory, and the cans were only partly filled, but there was no leak. 
A filling fault is very unlikely for Coca Cola or Pepsi, for which the Quality Assurance will be very high, but for a cheaper imitation brand it is possible. Bottling workers LAD and Mikeincharleston and canning manager HeadyCanner give answers in this thread regarding the likelihood of faulty seams causing leaks. They do not mention the possibility of under-filling, so I guess this is rare. It is also very unlikely that 2 cans from different "lines" would be underfilled.
Before opening you could submerge the cans in warm water and look for bubbles, which will confirm the leak theory - but will not tell you how long ago it started. If that doesn't work chill in the 'fridge for a few hours then submerge in hot water, again looking for bubbles.
Opening the can should give you more clues. If the liquid inside is thick and sticky and flat, this would confirm the leak theory - and indicate that it started long ago. If the liquid is normal fizzy coke, this confirms the filling-fault theory.
A: First, get someone else to open them :)
Based on Soda Cans
Check the use by date first, soda cans do not need to be absolutely leak-proof, they only have to be leakproof until the use-by date. 
Look at the top of the can, which is just a thin shell that the walls are crimped around. That's the only seal other than the ringpull,  over a long time span the carbonation can escape, a few molecules at a time. And with temperature changes water vapor inside the can will also escape. 
Coca Cola types drinks contain phosphoric acid, citric acid, and the carbonation is carbonic acid. These all attack metal, the walls and bottom of the metal containers are plastic lined  to reduce  the action of the acids on the metal, but it's not totally impervious. 
The only seals that are impervious to leaks are called hermetic seals, and these are not used for containers that contain beverages or most foods.
