Anthropic projects three AI scenarios for 2030, with extreme case lifting GDP 32.4%

September 9, 2026 at 05:00 AM2 min

This digest was compiled by AI from multiple sources — links to the originals are below.

The Anthropic Institute released a framework projecting how AI could reshape the US economy by 2030, with scenarios ranging from modest gains to 32.4% GDP growth. The model, reviewed by economists Daron Acemoglu and David Autor, links outcomes to AI's ability to perform knowledge work and its adoption speed. The extreme scenario would require recursively self-improving AI and could push cognitive-worker wages down more than 10%.

Key Facts

The Anthropic Institute's model projects US GDP growth of 1.6% to $34.1 trillion by 2030 in the modest scenario.
In the substantial scenario, AI could perform half of all knowledge work by 2030, lifting GDP by 8.3% to $36.3 trillion.
The extreme scenario, requiring recursively self-improving AI, projects GDP growth of 32.4% to $44.4 trillion and cognitive-worker wages falling more than 10%.
The framework was developed by a team led by economists Anton Korinek and Chad Jones and reviewed by Daron Acemoglu and David Autor.
The model estimates the US economy created $30 trillion in value over the past year across all tasks performed by people, machines, and software.

The Three Scenarios

The Anthropic Institute's framework divides the economy into tasks within occupations and examines what happens when AI improves, replaces, leaves unchanged, or creates new tasks. In the modest scenario, AI adds less than half a percentage point to GDP growth by 2030 and raises unemployment by only a tenth of a point, an impact similar to the internet. The substantial scenario assumes AI can perform half of all knowledge work by 2030, with most tasks still completed without AI, leading to GDP growth at twice the normal rate. The extreme scenario, which requires recursively self-improving AI, would double the economy every 4.5 years and push cognitive-worker wages down more than 10%.

Distributional Effects

Currently, about 60 cents of every dollar earned in the US economy goes to labor and 40 cents to capital. In the substantial scenario, wages for knowledge workers would not rise even as non-knowledge workers see gains. The researchers note that reallocation in the substantial scenario could be costly but is in line with what the US labor market has historically absorbed. The extreme scenario would see cognitive-worker wages fall more than 10%, while the economy as a whole grows rapidly.