Does Moore's law state that transistor counts double every two years without significant cost increase?
NoParrot Fact Check · May 20, 2026
NoParrot Score
Mixed signals
Claim breakdown
- Verified
- 0
- Uncertain
- 1
- Disputed
- 0
Synthesis
Moore's Law is widely understood to state that the number of transistors on a microchip doubles approximately every two years, with the cost per transistor remaining stable or decreasing. This is an empirical industry trend or projection, not a physical law, and the "without significant cost increase" aspect reflects the economic idea that computing density improves while the cost per transistor stays flat or falls. However, there is disagreement among models regarding the precise original formulation and subsequent revisions of this trend. Gordon Moore's initial 1965 observation suggested that component counts would double about every year, specifically at minimum cost per component. He later revised this interval to about every two years in 1975. Some sources indicate that the popular "every 18 months" version of the law factored in performance improvements alongside transistor count, as stated by Intel executive David House. Physical and economic limits are reportedly being approached, contributing to a slowdown in the trend in recent years. The core concept assumes that the doubling of transistors occurs without a proportional rise in cost per transistor, or at roughly constant cost for the improved density.
What each model said
- Gemini
Yes, Moore's Law states that the number of transistors on a microchip doubles approximately every two years, while the cost per transistor decreases.
- GPT
Yes, broadly. Moore’s law is commonly stated as: **the number of transistors on an integrated circuit doubles roughly every two years, while the cost per transistor does not significantly increase**—and historically often decreased. More precisely: - Gordon Moore originally predicted in 1965 that component counts would double about every year. - He later revised the pace to about every two years. - The “without significant cost increase” part reflects the economic idea that computing density improves while cost per transistor stays flat or falls. It is not a physical law; it is an empirical in…
- Claude
Essentially yes, though the precise wording matters. Gordon Moore's 1965 paper observed that the number of components per integrated circuit was doubling roughly every year **at minimum cost per component**. In 1975 he revised the doubling period to about **every two years**. The "minimum cost" aspect is integral to the law—it's not just about transistor density, but about the most economically optimal density. The popular "every 18 months" version came from Intel executive David House, who factored in performance improvements alongside transistor count. So your phrasing is largely accurate: M…
- Grok
Yes. Moore's original 1965 formulation observed that the number of transistors per integrated circuit (at minimum component cost) roughly doubles every year; he revised the interval to two years in 1975. The cost element is integral: the trend assumes this doubling occurs without a proportional rise in cost per transistor (or at roughly constant cost for the improved density).
Models Used
- Gemini
- GPT
- Claude
- Grok
Disclaimer: This page summarizes the output of 4 frontier large-language models at a single point in time. NoParrot Score reflects cross-model consensus, not absolute truth. Models can be wrong, and consensus can be wrong with them. Treat this page as a research aid, not a final verdict.
Related Questions
Check your own claim
Send any question to 4 AI models at once. See where they agree and where they don't — with a NoParrot Score from 0 to 100 and claim-level highlights.
Try NoParrot Free