Cognitive Scaffold

Preparing your thinking workspace

arrow_back_ios_new
MENTAL MODEL · M4950

Huang's Law

Huang's Law
TechnicalHigh supportArtificial Intelligence
Included
account_tree

Version 1.0.0 · Updated 2026-07-30

CORE DEFINITION

Describes the law of AI computing power growth: GPU inference performance doubles every two years, far exceeding Moore's Law. This benefits from software-hardware co-optimization (architecture, interconnection, algorithms), not just process advancement.

SCAFFOLDING EFFECT

psychology

Reduce cognitive load

The growth engine of the new era. If Moore's Law defined the era of general-purpose computing, Huang's Law defines the AI era. This means that any business relying on AI will see costs drop and capabilities improve much faster than traditional IT businesses.

anchor

Anchor fast decisions

An empirical law proposed by Jensen Huang of NVIDIA, stating that accelerators like GPUs, due to architecture, interconnection, and software co-optimization, achieve AI computing power growth far exceeding Moore's Law (about a thousandfold every 10 years).

MINIMUM ACTION

In progress 0/2

Practice this model in one real situation:

Check to track your progress (stored locally)
Learning progress0%
account_treeGenealogyexpand_more
menu_bookReferencesexpand_more

Source support: Explicit

  • link
    en.wikipedia.orghttps://en.wikipedia.org/wiki/Huang%27s_lawZH · Explicit
    verified

RELATED MODELS

Related model

Wirth's Law

Wirth's law is an adage on computer performance which states that software is getting slower more rapidly than hardware is becoming faster.

arrow_forwardRelated model

Vanishing Gradient

Very small backpropagated gradients can slow or stall learning in earlier layers or across long time intervals.

arrow_forwardRelated model

Allee Effect

Positive density dependence in which average individual fitness or per-capita population growth increases with population size or density over a given range.

arrow_forward