“No”, to your question about storage capacity of SSDs, the reverse if anything. No the fact that SSDs are small just comes down to cost, magnetic storage is very cheap, disks have several platters and bits per unit area figures just keep getting better and better. The cost per GB of SSDs will come down, and bit density will certainly go up. I assume that the price that the market is willing to pay is the limiting factor first. As well as that then I don't know if the physical size of the package is a second limiting factor assuming there are some users with really deep pockets. For very high capacity SSDs, lack of economies of scale will make things even worse as the demand for the entire package will be less healthy, and if they use higher bit-density chips then these may be less economical for various reasons for all I know, which doesn't help either.
Another thing with large SSDs. Why buy one when you could buy two units? (You might save money, or the reverse, I haven't looked. I ought to check which way the economic go.) But you will presumably get superb i/o bandwidth if you use RAID, a sexy combination, especially given the unimpressive read and especially write rates of some SSDs. (I would want a battery-backed RAID controller that is guaranteed to write back its internal RAM cache contents to the drives in a timely fashion, otherwise some of the reliability of SSDs is defeated.) And using raid striping with SSDs means you don't feel bad about such a loss of reliability as with mechanical HDs given they are so good to begin with. But anyway, it occurs to me that the option of running two SSD drives in parallel, even without RAID, might not help sales of the highest capacity units at all.
I ought to know more about SSDs' lifetimes. I've heard so much contradictory stuff, much of it vague, and little science. Operating systems matter - an SSD-aware file system is good for performance and a huge factor in keeping lifetimes healthy. Presumably very lazy write-back RAM caching helps a lot, although dangerous unless done well, but a lack of a correctly designed FS still means that there can be hotspots on the volume which are abused in relative terms. I wonder if a controller could fix the problem, by using an internal block mapping table, but of course this has to be done cleverly so that the region to which the table is eventually committed doesn't become a hotspot in itself. This unsolicited rant (apol) is off-topic (apol again) and might be worth splitting off into a new thread, but only if anyone might be interested in educating me / discussing SSDs. Admins?