US AI Economy Grows 2600% Annually but Invisible in GDP Statistics
Anthropic
Economists from the University of Virginia, Anthropic, and the Bank of Canada find that the US AI economy is growing at roughly 2,600% per year in quality-adjusted real terms, yet this growth is largely invisible in conventional GDP statistics. The gap arises because AI inference prices fall almost as fast as quality-adjusted output rises, making nominal revenues grow only moderately.
Researchers from the University of Virginia, Anthropic, and the Bank of Canada have published a paper showing that the US AI economy is expanding at an unprecedented rate, estimated at around $250 billion in nominal AI GDP for 2025, with quality-adjusted real growth of roughly 2,600% per year. However, this surge is not reflected in standard GDP measures because the rapid drop in per-unit prices for AI capability nearly offsets the rise in quality-adjusted output. The paper highlights that US compute spending on AI rose from $37 billion in 2023 to $90 billion in 2024 and $219 billion in 2025, while raw compute capacity grew over 200% per year due to chip efficiency gains. The authors warn that policymakers relying on conventional data may underestimate the potential labor market shock from AI substituting human labor, and recommend creating AI satellite accounts, improving data collection, and incorporating AI productive-capacity measurements into medium-term economic projections.
- Сокращения
- GDP = Gross Domestic Product — валовой внутренний продукт
Source: Import AI —
original