Anthropic Economic Index June 2026 Report

As Claude's usage patterns evolve beyond simple conversation into long-term Agentic Task execution through Claude Code and Cowork, Anthropic has upgraded its data pipeline to analyze economic indicators.

The new analysis method introduces fine-grained hourly data sampling and a new classifier for categorizing conversation purposes. Through this, data from Claude Conversations (including Chat and Cowork) and the 1P API is separated on a monthly basis to measure economic ripple effects with greater precision.

Key findings from the analysis are as follows:

• Cyclical usage patterns: Claude usage reflects work cycles. On weekdays, business tasks dominate, while on weekends, personal use (emotional support, medical questions, etc.) surges to about 50%. Time-of-day patterns also emerge, with news queries in the morning and sleep-related advice around 5 a.m.
• Output value and compute resources: Higher-value outputs generated by users tend to correlate with increased Token consumption and compute resources. Complex tasks like website building delegate more judgment to Claude than tasks like document translation.
• User expectations: The more users employ Claude in automated ways, the more they expect AI to take on greater workloads in the future, leading to optimistic outlooks that this will bring positive changes in terms of salary, job security, and meaning of work.