Since about 1980 we’ve gotten used to a ridiculous pace of technological innovation. At least as concerns processing power, it’s likely we just can’t keep up this pace of advance without a near-reinvention of the modern CPU. Chew on this: between then and 2005, the speed of available PC processors increased roughly 1000x.
Just compare that to other “indicators of progress” in our world: we definitely didn’t get one thousand times healthier (~0.29x fewer US deaths / 100k population ’80-’85 per the CDC), wealthier (~1.75x greater mean US net worth ’04 vs. ’83 per the Atlantic) or more numerous (world population ’80-‘05 via Wolfram Alpha: ~1.46x) over that period. Personal computers and the smart consumer electronics that followed have all benefitted from this Golden Era of processing power expansion. Moore’s Law—the predicted annual doubling of transistor counts—has become an expectation.
Unfortunately, Moore’s Law, which has already run into some bumps in the road, is widely expected to break down. But the good fight continues. DARPA, for example, recently launched an initiative to build a sub-20MW exascale computer, (1000 petaflops vs. our current max of about 50. Semis are by no means my area of expertise but I’m guessing there will be some interesting opportunities for innovation here. In any case, DARPA has done a few cool things in the past, like creating the internet, but my money would be on either a startup or a large consortium of tech companies (polar opposites, but hey, this stuff takes really big and patient money). A point for context: Q1 2014 saw 36 deals in Chips / Semis, the most in 2 years (representing $270MM in capital invested). For the sake of the innovation hungry consumer, good luck to them all!
PS: As I mentioned, this isn’t my area of expertise. But I’d love to hear anyone’s thoughts on which areas of innovation might be interesting here. Give me a shout: @picnoulos.