Blog Network

Anthropic · 2026-09-09 · major

Econ Scenario Explorer — Anthropic models three AI futures for 2030

Anthropic's Econ Scenario Explorer is an interactive model of the US economy in 2030 under three AI paths. Its extreme case puts GDP 32.4% higher at $44.4T while knowledge-worker pay falls more than 10%.

Cover graphic for the Anthropic Institute's Economic Scenarios for Transformative AI

Anthropic published an interactive model of how AI could reshape US growth, jobs and wages by 2030.

Quick facts

PublisherThe Anthropic Institute
What shippedEcon Scenario Explorer v1.0 plus a technical report
HorizonThe US economy in 2030
Modest scenario+1.6% GDP ($34.1T), unemployment in its historical range
Substantial scenario+8.3% GDP ($36.3T), unemployment around 5%
Extreme scenario+32.4% GDP ($44.4T), knowledge-worker pay down over 10%
Survey baseMore than 10,000 Americans

What is it?

The Econ Scenario Explorer asks you two questions — how much of knowledge work AI can do, and how fast firms actually adopt it — and then shows what the US economy looks like in 2030. The Anthropic Institute published version 1.0 on September 9 with a technical report, Economic Scenarios for Transformative AI, and survey data from more than 10,000 Americans. Three preset paths, modest, substantial and extreme, mark out the range.

How does it work?

Capability and adoption are the two dials; GDP, unemployment and wages by occupation are the outputs. The modest path lands at +1.6% GDP by 2030 ($34.1T), roughly what the internet did. The substantial path reaches +8.3% ($36.3T) with unemployment near 5%. The extreme path reaches +32.4% ($44.4T), and there knowledge-worker unemployment rises past recession levels while their wages drop more than 10%.

Why does it matter?

Every path the model produces sends a bigger slice of output to capital instead of to workers — capital's share climbs from about 40 cents on the dollar today to between 40.6% and 54.8%. That turns Anthropic's central claim into a distribution question rather than a growth question, and it hands policy teams a shared set of numbers to argue over instead of competing intuitions.

Who is it for?

policy researchers, economists and teams planning around AI's effect on work

Frequently asked questions

Who built Anthropic's Econ Scenario Explorer?
The Econ Scenario Explorer and its technical report, Economic Scenarios for Transformative AI, were developed by Anton Korinek, Chad Jones, Szymon Sacher, Tess Cotter and Peter McCrory. Santi Ruiz wrote the accompanying piece with that team. The work is published by the Anthropic Institute, the research arm Anthropic uses for its economic and societal analysis.
Can you change the assumptions in the Econ Scenario Explorer?
Yes. The Econ Scenario Explorer takes your own predictions about how capable AI becomes and how widely it gets adopted, then reports the resulting 2030 figures for growth, jobs and wages. The three named scenarios — modest, substantial and extreme — are preset starting points rather than the only answers the model will give you.
What do Anthropic's scenarios say about who gets the money?
Today about 60 cents of every dollar the US economy produces goes to workers and 40 cents to capital. Across the Econ Scenario Explorer's paths, capital's share rises — from 40.6% in the modest scenario up to 54.8% in the extreme one. Anthropic's report notes the gains in the extreme scenario are distributed unevenly even as total output grows.
How bad do jobs get in the extreme scenario?
In the Econ Scenario Explorer's extreme scenario, AI handles nearly all knowledge work, GDP reaches $44.4T, and unemployment among knowledge workers spikes past levels typical of a recession while their wages fall by more than 10%. Workers outside knowledge work fare better in that path than knowledge workers do.

Try it

https://www.anthropic.com/institute/econ-scenarios

Sources · 3 outlets

Tags

  • anthropic
  • anthropic-institute
  • economics
  • labor-market
  • ai-and-jobs
  • policy
  • forecasting
  • research-report
  • interactive-tool

← All releases