Cursor subscribers paid for Claude Max more often than comparison cohorts
Share with a direct Anthropic receipt in the $100–$200 Claude Max price bands, June 2025–August 2026
Share paying for Claude Max
Source: YipitData U.S. e-receipt panel. Cursor n=3,891; non-Cursor GitHub Copilot n=9,272; non-Cursor and non-Copilot ChatGPT n=29,552. Cohorts were established and equally observable.
Chart data
| Series | Share paying for Claude Max |
|---|---|
| Cursor subscribers | 21.4% |
| GitHub Copilot subscribers | 11.1% |
| ChatGPT subscribers | 4.9% |
How often did the cohorts pay Anthropic directly at any consumer tier?
Direct Anthropic receipts appeared for 57.2% of Cursor subscribers, 38.8% of non-Cursor GitHub Copilot subscribers, and 28.2% of the ChatGPT comparison cohort. Within Cursor, 34.6% paid at the $20 Claude Pro price point and 21.4% paid at the $100–$200 Claude Max price points.
Was the premium overlap specific to Claude?
Within the Cursor cohort, 21.4% had a Claude Max receipt while 0.5% had a $200 ChatGPT Pro receipt. The difference shows that the measured premium-tier overlap was concentrated in Claude, but receipt data does not reveal which tool carried the primary workload or whether users treated the products as substitutes.
How did Claude Max co-holding change within fixed Cursor and Copilot cohorts?
Among 1,710 Cursor subscribers present before June 2025, Claude Max co-holding rose from 1.4% in the first quarter of 2025 to 19.9% in the second quarter of 2026. A fixed cohort of 3,752 GitHub Copilot subscribers rose from 0.4% to 8.0%. The parallel increase does not isolate the effect of Cursor pricing or establish future churn.
Assumptions & Methodologies
We analyzed established U.S. e-receipt panelists active from June 2025 through August 2026. Cursor used direct merchant receipts. GitHub Copilot was isolated from GitHub line-item text and excluded Cursor subscribers. Claude Pro and Max used direct Anthropic USD receipts at $20 and $100–$200, excluding irregular API charges. The fixed-cohort trend followed pre-June 2025 Cursor and Copilot subscribers quarterly. Cursor tier and usage were unobserved, Anthropic receipt capture changed over time, and cohort differences can reflect selection. Results measure co-payment, not usage, substitution, churn, or pricing causality.