Moore's Law

(Work in Progress) "Moore's Law is dead" - Gordon Moore Moore’s law is no longer expected to deliver improved transistor cost scaling at or below the 20nm node. Up until now, the cost per transistor of a new technology has always been cheaper. Now we face a scenario in which the cost of new transistors will be more expensive. That means more transistors = bad, which puts profound pressure on the R&D pipeline in different ways. The other drivers of Moore's Law -- clock speed and Instructions per Clock -- plateaued a long time ago. New Technologies It takes 15-20 years to move tech from theory to consumer hardware. There's a lot of theory out there, and no commercialized replacements. We are approaching the point where feature sizes can be described by atomic widths. There is no room to get much smaller or more perfect. Graphene is decades away. So are carbon nanotubes. SiGe doesn't scale well; InGaAs is fragile and has poor manufacturing characteristics, a host of other problems affect every other alternate substrate. Optical computers. Impact Computers are expected to be 30x faster in 50 years. Whatever alternate means of computing we use (spin, magnetic charge, measuring electron positions) will likely be confined to huge labs and expensive installations. You can do quantum computing today, but not without a hefty supply of liquid nitrogen.