It has been over ten years since I first heard of someone claiming that their version of nanotechnology would ‘replace’ CMOS but there still isn’t a challenger in sight. Today’s announcement of Intel’s 3D finFET technology illustrates just how good the semiconductor industry is at controlling materials on the nanoscale, and there are plenty of tricks left in the bag yet.
The problem for any challenger is that after forty years, the semiconductor industry can make a few silicon atoms do pretty much what they want, hook them up with other components and then connect them to the outside world. Even better, using the accumulated design and modelling expertise you can predict almost exactly what any given process will result in, even down to the yield and design a production process to do just that. And as long as people want to buy semiconductors, the industry can continue to reinvest some of its revenue in R&D at a level that no start up, not matter how well backed, could ever hope to match.
Any new material, such as graphene, carbon nanotubes or organic electronics simply doesn’t have that depth of knowledge about processing and manufacturing so “replacing” CMOS just isn’t an option in the near future.