In 1965, Gordon Moore, then head of Research and Development at Fairchild Semiconductor, prepared an article for the 35^th anniversary of Electronics Magazine. He discovered that over the period 1959-65, the number of components (e.g., transistors) on a chip roughly doubled every year.

Figure 1, reproduced from Moore's 1965 Electronics Magazine article, shows his findings. This trend suggested that processing power would rise exponentially and at a fast rate, leading to a computing revolution. Moore predicted: "Integrated circuits will lead to such wonders as home computers - or at least terminals connected to a central computer, automatic controls for automobiles, and personal portable communications equipment." This is a remarkable prediction given that it was made in 1965.

Figure 1: Moore's Original Law: Logarithm of the Number of Components on a Memory Chip Over Time.

^{In the 1975 IEEE International Electron Devices meeting, Moore revised his prediction to the number of transistors on a chip doubling every two years.'
Figure 2 shows the growth in the number of transistors in Intel's microprocessor chips from 1971, when the first (Intel 4004) microprocessor was introduced, through the introduction of Intel's Pentium II in 1997. The number of transistors on a chip almost doubled every two years. Similar results hold for the number of transistors on a DRAM chip, which have in fact increased at a somewhat faster rate. Thus, Moore's 1975 prediction is generally consistent with the empirical evidence.}

¹

Most newspaper accounts cite Moore as saying the number doubles every 18 months. This is plainly inconsistent with history. In fact, the regression equation shows a growth factor of 1.9 in two years, with R²in excess of 99%.

2

With the increase in chip density came a corresponding decline in the costs of computation. As these costs continued to decline exponentially over a period of thirty years, the compound effect was powerful: when costs are cut in half every two years, over a 30-year period they decline by a factor of 2¹⁵, turning a million-dollar processor into a $30 device. The result of these powerful trends was the computer revolution, and with it - the information revolution. Lowering the costs of processing and storage enabled the development of powerful software that took advantage of the ever-increasing capabilities of the hardware. The new software gave rise to new applications that in turn increased the demand for hardware. As Moore's Law progressively reduced the costs of computation, additional applications crossed the threshold of affordability, further increasing the demand for computing. Thus, Moore's Law was the driver of a positive feedback loop that made Gordon Moore's 1965 predictions come true: the integrated circuit is playing a central role in every facet of modern life. Indeed, Moore's Law is driving the acceleration of "clockspeed" in the entire Information Technology industry and beyond. Moore's Law was not only a prediction - it also formed a blueprint for the development of the semiconductor industry.

Since Intel was founded in 1969 by Robert Noyce and Gordon Moore, Moore's Law became a target that drove product development within the company. In fact, the entire semiconductor industry is striving to track Moore's curve: the Semiconductor Industry Association puts together periodic "Technology Roadmaps" that were closely followed by the chip industry. These roadmaps, designed by technology working groups made up of leading industry experts, define in detail the course for future developments over a 15-year period, driven by the desire to continue the past trends of Moore's Law. In this way, Moore's Law has become a self-fulfilling prophecy.