Lack of GPUs prolongs profitability for those who already have them, or manage to get some.
However, do keep in mind that NVidia makes GPUs too, and they are very profitable on most coins and can often beat AMD on hash/watt even where they lose on hash/$ (mostly ETH ETC and such, and the Cryptonight coins like XMR).
For ZEC (which isn't the most profitable NVidia coin by far but is probably has the second most GPUs mining on it) NVidia is quite close on hash/$ (and might even WIN on some comparisons) WITHOUT FACTORING IN the current AMD "gouge" pricing.
Funny part - folks have been pushing the Skein coins a lot, but I actually ended up making more on something else (that I'm NOT going to mention here, as it's fairly small and I don't want the profitability of it to get overwhelmed by a bunch of newbies or even one BIG farm jumping all over it).
Mentioning it or not, it doesn't really matter in the end because it eventually evens out with other coins profitability, lots of people know about whattomine, and also people can check what's hot on trading volume on any exchange especially poloniex which is booming hard... probably you meant SIA,(but it doesn't matter)
SIA went up by a lot last couple of weeks and I jumped on the hype train just today, dual mining eth has finally become more profitable for me...
But Sia is just one of many, generally every altcoin has been pumped since last ~2 months...
Whattomine doesn't list all algos on the compare page, so yes it DOES matter.
9-)
For one example, Skein isn't there (even though at least one Skein-based coin is TON more profitable than Pascal which IS there).
Additionally, the "hashrate" shown on many algos on many cards is WAY off.
There are already quite a few coins that NVidia cards mine for higher profitability than AMD cards do - even ZEC in the last month or so has been seeing NVidia achieve equal-to-better hash/$ vs AMD, and ZEC isn't exactly small based on the number of cards needed to achieve the current network hash (I think ETH is the only "bigger" GPU-mineable coin on that basis).
I doubt that profitability of all coins will return to where they were at a year ago - but I DO expect to see a shakeout and a major dump on profitability for most GPU-based coins over the next year or two - probably triggered when ETH starts *finally* phasing out POW heavily (ballpark when it moves POW to half or less of total "Proofing").
The interesting thing is if that does end up being the timing, it most likely happen pretty close to the same timeframe that the next generation of SHA256 (and probably Scrypt and possibly X11/X13/X15/etc) ASIC miners start selling.
Moore's Law has been taking a noticeable hit over the last decade, as current semi tech runs seriously into quantum-level issues, and Silicon reaches the "end of the road" as a viable semi material. The next decade could see Moore's Law end - and it looks definite that it's going to have a major PAUSE as a minimum after the 7/10nm nodes go into production.
! Plus we all thought GPU mining was dead, but Ethereum revived it two times. And yet that is sure there will not be a third time because of Casper, there will still be some coin a day or another, just as ZCash, that will pop up out of nowhere and will be profitable. 
