Moore's Law — Historical Trajectory and Where It Bent
From the 1965 Gordon Moore paper to FinFET, GAA, and the end of geometric scaling — what continues to scale and what doesn't
Gordon Moore's original 1965 paper is four pages long, and the famous prediction sits in a single paragraph. Moore — then Director of R&D at Fairchild Semiconductor, three years before he co-founded Intel — observed that the number of components on an integrated circuit had been doubling every year since the IC was first commercialized in 1959, and projected that pace would continue for at least the next ten years. He did not call it a law.
The phrase Moore's Law was coined by Carver Mead, then a Caltech professor, several years later as a marketing nickname. What Moore predicted — exponential doubling of transistor density — has held in some form for roughly 60 years, the longest empirical scaling pattern in industrial history. What he could not predict was the cost economics.
The doubling continued; the per-transistor cost reduction that originally accompanied it stopped sometime around the 28-nanometer node, roughly 2012-2014.
That is the opening. Finishing a lesson is where it stops being interesting and starts being useful: the full lesson runs to 7 sections and ends with 6 practice questions. A free account is what opens the rest, and the other 255 lessons in the Academy with it. No card.
What this lesson covers
- 1Dennard scaling — the physical engine that powered Moore's Law, and why it broke
- 2What is still scaling and what isn't
- 3Major process node introductions, 1971-2025 — what each generation brought
- 4Cost-per-transistor scaling — the part of Moore's Law that bent
- 5Intel 22nm Tri-Gate, May 2011 — the architectural rescue that bought the industry a decade
- 6Where to see this on the platform
- 7Summary