Hot Chips keynote speaker Robert Colwell (director of the microsystems group for DARPA) has predicted the end of Moore’s Law possibly by 2020, with process technologies reaching their limits around the 7nm node. In particular, the limiting factor—rather than physics per se—will be financial, according to Colwell: “When Moore’s Law stops it will be economics that stops it, not physics, so keep your eye on the money.”
Making firm predictions regarding the end of Moore’s Law has made many a false prophet. A growing consensus suggests that it will end before too many more decades, but the exact departure point—as well as what that would look like precisely—is murkier. More-incremental improvements in processors (and a greater focus on software) will likely follow the demise of exponential processor development, pending a major (albeit unlikely) breakthrough in quantum computing or another area on the frontiers of research. Either way, far more fundamental to Moore’s Law than physics is the fuzzier field of economics.
Read more about Moore’s Law