Anthropic models AI economy through 2030

Anthropic has published a new economic model that sketches three ways artificial intelligence could reshape the US economy by 2030, ranging from a modest productivity boost to a boom paired with white-collar job losses. According to Euronews, the company’s “extreme” scenario shows annual growth reaching 15.4% while unemployment among office workers rises to 17.9%, with overall unemployment at 11.9%. The same scenario has labor’s share of national income falling from 60% to 45.2% as capital income rises sharply.

The work is framed as a set of scenarios rather than a forecast. Anthropic’s economics team released a technical paper and an interactive tool that model the economy as bundles of tasks that AI can leave untouched, assist with, automate, or create. In the “substantial” scenario, AI performs about half of knowledge work and the economy’s growth rate roughly doubles, while cognitive employment falls and office-worker unemployment rises to 4.5%. In the “modest” case, the model shows little movement: GDP in 2030 sits only slightly above a no-AI baseline and unemployment barely changes.

What stands out is not just the spread between scenarios, but where the model places the gains. In the high-growth case, office wages fall while wages for other workers rise, and the biggest shift is from labor income to capital income—an outcome that depends less on whether AI is impressive than on who owns the systems and the complementary assets around them. Anthropic’s own write-up ties outcomes to adoption speed as much as technical capability, a distinction that matters because diffusion is governed by procurement cycles, regulation, and corporate risk tolerance, not demo videos.

The release also lands amid sharper internal rhetoric from the same industry. Euronews notes that Jacob Coxon, a researcher who worked at OpenAI and Anthropic, resigned and criticized what he described as a race toward self-improving systems, arguing that executives soften public language compared with private fears. Anthropic’s model does not include scenarios where AI causes severe negative outcomes; it assumes the economy continues to function, even in the most disruptive labor-market case.

Anthropic is asking readers to explore an interactive tool showing office unemployment near 18% in one pathway, while its own scenarios stop short of modeling what happens if the institutions managing that transition fail.