AI Estimated to Save $2.7 Trillion in Labor Costs Annually

A study has revealed that the cost equivalent of labor hours saved by AI amounts to approximately $2.7 trillion annually. Researchers analyzed the time-saving effects of AI, which are spreading beyond software development into fields such as administration, education, and sales. The study was conducted by analyzing five iterations of the Anthropic Economic Index from January 2025 to February 2026. By comparing wage and employment data from 86 countries, it is estimated that the labor cost equivalent (LCE) of time saved by AI is $2.7 trillion, which corresponds to 3.4% of the GDP of the surveyed countries. Six months ago, the estimated LCE was $1.2 trillion, indicating that this estimate has more than doubled. LCE serves as an indicator of the productivity value of AI usage, with variations depending on wage levels. The spread of AI usage has expanded from initial software development to education, administration, and sales, with average wages decreasing by 5.5% over 13 months. However, the overall labor cost equivalent has increased. The LCE for high-income countries is 4.2% of GDP, significantly higher than 0.6% for middle-income countries and 0.1% for low-income countries. The AI Concentration Index (ACI) measures the distribution of AI usage value, with the ACI for the United States recorded at 0.44. Researchers emphasized the need to consider both the scale of saved hours and the distribution of their value to assess the economic impact of AI.