However, we just got our own instrument, so this will definitely be an issue, but our University knows a thing or two about dealing with big data (http://kb.iu.edu/data/avvh.html).
I've only dealt with a few GWAS style datasets, and the next-gen stuff dwarfs the GWAS data in terms of size. But when looking for linkages between variations, we're still talking more time than the universe is old level of calculations for more than 3 combinations. Which is really scary, because like you said, all the genetics people are going to be using sequencing for most things from here on out, so its like you have complexity on top of complexity...
OK, now that I've thought about it, the easiest way around this probably is throwing money at hardware (more disks) or optimizing the processing. However, this is only in the case of a true disk I/O bottleneck. If you're optimizing correctly then the disks should be reading 8GB blocks directly into memory and the CPU should be spitting them right out again. At the very minimum you should be using an optimized file system with large pages enabled in your kernel.
"What do you do for a living?"
"Exomes."
-spits drink out-
Where I work, filesystem I/O is the rate-limiting step for most of my (next-gen sequencing) experiments...