After 44 years and surviving several prediction of its demise, is Moore's law really dying? Pundits are saying that there is a limit to innovation & this time, Moore's law has met its nemesis. Unfortunately, its not the laws of Physics. Its the "Green Backs", money? At a certain point, the cost of producing smaller chips will cross the line of being economical.
So what is Moore's law - its a prediction or a trend that every two years the number of transistors will double on an economical Integrated Circuit. Pundits are saying that the limit is 18nm feature size, the smallest feature that can be made on a chip. We will get there by 2014 or 50 years from the day Moore's law was established by Intel's Gordon Moore.
Over four decades, technological innovation and insatiable demand for chips delivered on the Moore's law predictions. While the cost of setting up a chip manufacturing plant, Fab has sky-rocketed to over $4B, the chip demand has kept pace, providing the economies of scale. Today, chips are ubiquitous. Every time we felt an imminent inflection point, a new usage of chips came through - PCs and Cell phones in last 2 decades.
The survival of the Moore's law depends upon finding new usage? Is it smart phones, network PCs, electronics book, or something entirely different? Perhaps, a different kind of solar cell or energy storage device or nano devices in healthcare. Its the unpredictable nature of the innovation that has made Moore's law so predictable.
Subscribe to:
Post Comments (Atom)
While I agree that Moore's law has done an impressive job forcasting integrated circuit development, I don't believe unpredictability(volatility) is the source causing the path of progression. Weather variations are an example of unpredictable effects which doesn't cause an exponential effect.
ReplyDeleteRather, I think it's the diversity of sources of innovation which has led to the predictable exponential path estimated by Moore's law. Each potential innovate new use of the integrated circuit is unpredictable, but the many avenues of research and commercialization of new electronic products leads to predictable aggregate innovation. Much like you can't expect the dividend of one specific company to continually increase, a regular annual dividend increase is much more likely holding a broad stock index.
-Andrew Huang