Moore's Law is dead, but innovation will continue to be

The "silicon area", that the base of the production of semiconductors present in our computers, is no longer able to take the step established by the famous "Moore's Law", with the theory that is soon destined to be buried. This was stated by a substantial deepening of Nature.com site, which claims that soon the company will stop industry to pursue the dictates of the old theory, that only last year celebrated its fiftieth year of birth.

Moore's Law has been theorized for the first time by the legendary co-founder of Intel, Gordon E. Moore, in 1965. It was enunciated in various ways over the years, but in its simplest form the concept expressed by its author it is that "the complexity of a chip, measured for example by the number of transistors per chip", would be doubled from year to year. An assumption that the companies in the sector, for example, the same Intel, have managed to keep up to date considering the revision of 1970, with the theory that changed envisaging doubling every eighteen months.

The long study of Nature, however, tells a different reality for the coming years, a reality that the semiconductor industry is expected of time. Modern chips are already quite complex in their implementation and use production processes already extremely miniaturized, to the point that it is becoming increasingly complex and difficult to increase their density. In view of this it seems very difficult today to be able to double its capacity every 18 months.

The most advanced chips currently on the market using production processes at 14nm, where a nanometer is a billionth of a meter. At this size the individual parts that make up the chip are smaller than a typical viral particle, and the like to the size of the outer membranes of a germ. Print a 14nm chip has become so complex and expensive that to date only four companies manage to keep pace: Intel, Samsung, TSMC and GlobalFoundries.

In comparison, only 10 years ago there were no less than 18 on the market. The next step will follow the market will be that of 10nm, which should arrive in the chip for the consumer market in 2016. Later come the 7nm, expected in 2018 or in 2019, while TSMC has begun work on a manufacturing process with 7nm the goal of having the first prototypes ready for next year. It will be with 5nm that there will be the first issues of a physical, hard to get around with the current skills.