Dual subscribers were about nine times as likely to keep only ChatGPT
Status in June–August 2026 among panelists active on both Midjourney and ChatGPT in June–August 2025
Share of baseline dual subscribers
Source: YipitData U.S. e-receipt panel. Midjourney and ChatGPT dual-subscriber cohort n=426. Of the cohort, 203 dropped Midjourney but kept ChatGPT, while 22 dropped ChatGPT but kept Midjourney. The remaining 47.1% either kept both services or dropped both; the supplied evidence did not separate those outcomes.
Chart data
| Series | Share of baseline dual subscribers |
|---|---|
| Kept ChatGPT only | 47.7% |
| Kept Midjourney only | 5.2% |
Did the same pattern appear among Midjourney and Claude subscribers?
Yes. Among 180 panelists active on both Midjourney and Claude in June–August 2025, 47.8% dropped Midjourney but kept Claude a year later, while 4.4% dropped Claude but kept Midjourney. The smaller Claude cohort makes this a directional confirmation rather than a second precise estimate.
Assumptions & Methodologies
We analyzed U.S. e-receipt panelists active on both Midjourney and ChatGPT in June–August 2025 and measured status again in June–August 2026. Midjourney activity required a direct billing receipt; ChatGPT activity used the latest consumer plan or plan-will-not-renew signal. Developer invoices were excluded. Panelists were observable before baseline and active within 30 days. The Claude replication included 180 panelists. Receipt models differ and silent cancellations may be missed, so absolute gaps are directional. Dual subscribers are heavier AI spenders by construction, and Gemini and platform billing are excluded. The outcomes are not proof of causal switching.