Archive · 10 Sep 2026
Anthropic Projects AI-Driven GDP Growth Amid Looming Jobs Crisis
Anthropic made three simulations showing how AI might change the US economy.
These are not predictions, but examples based on different assumptions.
In the modest scenario, AI helps the economy about as much as the internet did.
In the substantial scenario, AI can perform about half of knowledge-work tasks by 2030.
In the extreme scenario, AI performs nearly all such tasks on its own.
This could make the economy much bigger and eventually lead to very fast growth.
However, some knowledge workers could lose tasks, earn less or need to retrain.
People who own AI systems and computers could receive a larger share of the money created.
The final outcome will depend on how capable AI becomes, how widely it is adopted and how its benefits are distributed.
Anthropic modeled modest, substantial and extreme AI scenarios for the US economy through 2030, stressing they are simulations rather than forecasts.
US GDP could reach $34.1 trillion, $36.3 trillion or $44.4 trillion by 2030, depending on the scenario.
In the extreme scenario, annual GDP growth could reach 15%, but unemployment could rise beyond typical recessionary levels.
Knowledge workers, including coders and customer-service workers, could face flat or falling wages as AI automates more tasks.
AI adoption could shift income from labour to owners of AI systems, computing infrastructure and other capital.
- Who
- Anthropic and more than 10,000 Americans surveyed by Anthropic.
- What
- Anthropic modeled three possible AI-driven economic futures and assessed their potential effects on GDP, jobs, wages and the division of income between labour and capital.
- When
- The scenarios focus mainly on outcomes by 2030; the report was published on September 10, 2026, and the extreme scenario includes longer-term growth estimates.
- Where
- The estimates concern the United States.
- Why
- The simulations were designed to show how different assumptions about AI capability, adoption and productivity could produce different economic outcomes.
Economic Expansion and New Demand
Job Disruption and Unequal Gains
Overall economic effect
Economic Expansion and New Demand
AI could raise productivity and make the US economy substantially larger, with GDP gains ranging from 1.6% to 32.4% by 2030 across Anthropic's simulations.
Job Disruption and Unequal Gains
A larger economy would not necessarily produce proportionally higher incomes for workers, especially those whose tasks are highly exposed to AI.
Impact on workers
Economic Expansion and New Demand
AI could augment existing jobs, create new responsibilities and increase demand for workers in less AI-exposed occupations, including construction workers.
Job Disruption and Unequal Gains
Coders, customer-service workers and other knowledge workers could lose tasks, need retraining, face flat wages or experience wage declines.
Distribution of gains
Economic Expansion and New Demand
Productivity improvements could increase returns to companies and individuals providing AI systems, computing infrastructure and other productive capital.
Job Disruption and Unequal Gains
A shift toward capital could reduce labour's share of GDP and concentrate more of the gains among capital owners.
- Scenario types
- Modest, substantial and extreme
- Modest 2030 GDP
- About $34.1 trillion, or 1.6% above an economy without AI
- Substantial 2030 GDP
- About $36.3 trillion, or 8.3% above an economy without AI
- Extreme 2030 GDP
- About $44.4 trillion, or 32.4% above an economy without AI
- Extreme growth
- Annual GDP growth could eventually reach 15%, doubling the economy roughly every 4.5 years
- Labour and capital
- In the extreme scenario, labour receives 45.2% of GDP while capital receives 54.8%, compared with an initial split of roughly 60% and 40%
- Survey expectations
- The typical respondent's answers implied an economy about 10% larger by 2030, with unemployment around 5%
