Now, the large number of samples/files set the limits for how much we can load. But...
When you create a gigastudio instrument, there is nothing stopping you from putting several samples into the same audio file. You can have several sounds running after each other in the sound file. It's not a big deal. The first sound runs for a couple of seconds, then the next, and so on.
You can use this audio file for several different keys/velocity layers (dimensions) but specify a different starting position (in the audio file) for each, to get only the sample you want.
It isn't possible to set an end position. But, if the sound is looped this isn't an issue, at least if it doesn't have a tail. If the sound isn't looped, just put a short loop at the end of the sound (a silent loop) to make sure the next sound in the audio file isn't played when first sound has finished.
Theoretically, you could store all the samples in just one audio file, and just specify different starting positions.
Now, an important question, if you have different starting positions, do gigastudio cache these as different samples in RAM? If it does so, this is pointless.
Another issue, if the first couple of seconds aren't always stored traditionally in RAM, will this be a performance issue? Maybe not, since lack of RAM really is the bottleneck now.