Anthropic released an AI economic scenario, predicting a 32.4% increase in the U.S. GDP by 2030 under extreme conditions

Anthropic's economic team builds models to analyze the impact of AI on employment, growth, and unemployment in the United States. The model views the economy as a combination of tasks based on the U.S. Department of Labor's O*NET classification, where AI can keep tasks unchanged, enhance, automate, or create new tasks. The U.S. economy is described as having a task instance value of over $30 trillion.

Three scenarios depend on AI capabilities and adoption speed: the moderate scenario is similar to the internet, with U.S. GDP increasing by 1.6% by 2030 ($34.1 trillion in 2025 prices), labor share at 59.4%, and capital share at 40.6%; in the substantial scenario, AI can perform half of knowledge work, GDP increases by 8.3% ($36.3 trillion), and wages for knowledge workers remain roughly stable, with a labor share of 56.1%; in the extreme scenario, AI is much more efficient in the vast majority of knowledge work and completes it almost entirely autonomously, with GDP increasing by 32.4% ($44.4 trillion), wages for knowledge workers dropping by over 10%, a labor share of 45.2%, and an unemployment rate exceeding typical recession levels.

A survey conducted in August with over 10,000 Americans showed that typical responses were close to the substantial scenario (2030 GDP about 10% higher, unemployment rate about 5%), with about 10% of respondents nearing the extreme scenario. GDP increases in all scenarios, and transformative scenarios require more job transitions. The page provides a technical report and an interactive explorer.