Yes, these "science" ideas (actually, traditional engineering) really are that hard to fund.
One thing software people always forget is how expensive it is to perform R&D on real, physical things. Software is practically free to produce in comparison - your big campus of 1000 engineers, with free lunch and masseuses, is practically free in comparison to say, the cost of developing a new drug, or introducing a revolutionary new jet engine.
I don't think it's necessarily true that VCs only care about iPhone/Android apps, but they honestly cannot afford to invest in "real" engineering. The risks are just as great, and the stakes are far, far higher, beyond the deep pockets of many angels and VCs.
I don't think you're the only one that's realizing this. I'm seeing more and more posts where people complain that startups aren't solving "real" problems.
Perhaps investors look at FB, Twitter, Google, LinkedIn, and think that these 4 statistical anomalies are examples to follow, and thus fund companies with similar ideas?
On the other hand, I have a hard time believing that the "real problems" aren't being addressed by anyone.
We have some very real problems in this world, but all you seem to care are iPhone/Android apps and stupid websites (copied from one another I guess).