http://en.wikipedia.org/wiki/DARPA_Grand_Challenge
http://en.wikipedia.org/wiki/Scanning_tunneling_microscope
http://en.wikipedia.org/wiki/Public-key_cryptography
http://en.wikipedia.org/wiki/Memristor
http://en.wikipedia.org/wiki/Fractals
All since 1970, is that recent enough? And, ya know, sometimes it does actually take a little time for major achievements to be recognized as such.
I'm not saying the things you list weren't hugely important, I'm just saying that they weren't as fundamental as the entirely new fields that were developed earlier (relativity, quantum mechanics, theory of computation, etc).
And you're being overly reductive when you say the Grand Challenge is a "pretty natural extension of Turing, Von Neumann, etc.'s". I'm pretty sure the physicists that you've mentioned would take exception to the idea that their work didn't have a clear lineage of its own.
I just feel like their were are far fewer (if any) of those truly ground breaking discoveries now than there were in the early 20th century. This might just be perspective bias, but it seems like scientific progress was being made much more quickly from, say, 1850-1950 than any other time in human history.
We still have a lot of big open questions, but progress seems to have slowed down. Hopefully the LHC will change that ...
As for the Grand Challenge, I work in the AI field myself (I'm engaged in some work for DARPA right now, and I know several people who were on Grand Challenge teams), and we're still basically stuck in turing/von-neumann paradigms. In my opinion, most of what we've done in the last 50 years in AI is come up with better search heuristics. It makes AI/ML much more useful - but it's nothing really paradigm shifting.
I will of course grant this could all be in my head, since I haven't, and probably couldn't, define a metric that measures the 'significance' of theory (maybe something about marginal information-theoretic predictive power, but that's difficult to measure and still not quite right).