Technically, the NSA have been doing what they are doing for a while. So irrespective of what you think about the NSA or Snowden, it was the information that they're doing it getting wider attention that is causing the demand to drop.
A tool to find exploits of open source software? That is so evil.
However, as justification for this opinion you give completely irrelevant comparisons.
I can go into the technical details if you insist as to _why_ it is the case, or you could go to to http://www.cs.ualberta.ca/~games/ and read around there (and contemplate why the University of Alberta considers their chess research a "previous project", while the Poker research is ongoing), or perhaps you'd like to actually implement either a chess engine or a poker bot or both, and then see how far you can get. Or you could read one of the actual scientific papers on Poker AI, where there's usually a justification why they think Poker research is more fruitful nowadays than chess research.
Alternatively, you could just trust me when I say:
The techniques used in writing a chess engine that can beat an average human player are so established that even beginning programmers can quickly implement them (drop by the forums at http://www.talkchess.com/forum/viewforum.php?f=7 for some non-academic discussions on chess engine writing). Essentially, just searching as many positions as possible as quickly as possible (and using the latest hardware) has been the most successful approach.
For Poker, just the fact that parts of the game are hidden makes it harder by orders of magnitude, and research is very active.
The number of combinations you can make with a deck of cards vs. the number of moves in a chess game is a completely invalid comparison. I'd suggest that any such comparison attempted would be more or less meaningless.
Just the fact that endless numbers of games playable with a standard deck of cards of a complexity ranging from the trivial to the extremely complicated should suggest to you that perhaps that the number of comparisons is perhaps not a very meaningful metric.
I'll pick out just one small factor to perhaps illustrate the huge differences involved:
Judging the strength of a chess engine properly is hard enough despite the fact that in principle the results are very easy to analyse: There is only either a win, a loss or a draw. Even with this superficial simplicity thousands of games have to be played before chess engine A can be confidently judged to be stronger than engine B. And this in a game where there's supposedly no random factor!
In Poker, even just determining whether one player was simply lucky or actually better is a research programme in its own right.
In the FA, a very limited variant of Poker was used: One-on-One Limit Texas Hold-em. It gets much more complicated once you add more players, and more complicated again when you go from Limit to No-Limit Poker.
Your "opinion" is wrong on so many levels that it takes real effort for me to stay on-topic and not flame you personally for it (I apologize if my effort was not sufficient).