Moore's law
/morz law/ · noun · Hacker Culture · Origin: 1965
Definitions
Gordon Moore's 1965 observation that the number of transistors on an integrated circuit was doubling roughly every year, revised by him in 1975 to every two years. It is an empirical trend and an industry planning target, not a law of physics, and its long persistence owed much to its becoming a self-fulfilling roadmap that semiconductor firms coordinated around. Moore's law describes transistor density specifically, and is frequently conflated with related but distinct trends in clock speed and performance; clock scaling effectively ended in the mid-2000s when power density made further frequency increases impractical, which is what pushed the industry toward multicore designs. Density gains have continued but have slowed markedly and grown far more expensive, prompting a shift toward specialised accelerators and advanced packaging.
In plain English: The observation that computer chips roughly double in power every two years — it's held true for over 50 years but is starting to slow down.
Example: Moore's law predicted this trajectory decades ago, but we're reaching the point where individual transistors are only a few atoms wide.
Etymology
- 1965
- Gordon Moore observed in Electronics Magazine that the number of transistors on integrated circuits doubled roughly every year, projecting this trend would continue for at least a decade.
- 1975
- Moore revised his prediction to a doubling every two years. This became the canonical formulation of 'Moore's Law,' driving semiconductor industry roadmaps.
- 1990s-2000s
- Moore's Law held remarkably steady for decades, guiding Intel, TSMC, and Samsung to ever-smaller process nodes. It became shorthand for exponential technological progress.
- 2010s-Present
- Physical limits of silicon transistors slowed traditional scaling. The industry shifted focus to architectural innovation (multi-core, chiplets, AI accelerators), prompting ongoing debate about whether Moore's Law is 'dead' or 'evolving.'
Origin Story
The Prediction That Powered Five Decades of Progress
Moore's Law is the observation that the number of transistors on an integrated circuit doubles approximately every two years, driving exponential improvements in computing power. Gordon Moore, co-founder of Intel, first made this observation in a 1965 article for Electronics Magazine, where he noted that the number of components on chips had been doubling each year since their invention and predicted this trend would continue. In 1975, he revised the doubling period to approximately every two years, and this adjusted prediction held remarkably well for nearly half a century. Moore's Law was never a physical law but rather an empirical observation and a self-fulfilling prophecy: the semiconductor industry used it as a planning target, with companies like Intel, AMD, and TSMC structuring their research and investment roadmaps around maintaining the pace. The economic implications were staggering. The exponential scaling meant that computing power became roughly a million times cheaper between 1965 and 2015. Smartphones with more processing power than 1990s supercomputers became affordable to billions of people. In recent years, the pace has slowed as transistors approach atomic scale, leading to debates about whether Moore's Law has ended or simply evolved into new forms of performance scaling.
Coined by: Gordon Moore
Context: First described in a 1965 Electronics Magazine article; revised to a two-year doubling period in 1975.
Fun fact: If the automotive industry had improved at the same rate as semiconductors under Moore's Law, a modern car would travel at roughly 300,000 miles per hour, get two million miles per gallon, and cost about four cents.