Poe subscribers paid for at least three AI tools at nearly nine times the comparison rate
Share holding at least three distinct paid AI tools from January 2024–August 2026
Share paying for at least three AI tools
Source: YipitData U.S. e-receipt panel. Direct-billed Poe subscribers n=155; other paid-AI subscribers n=104,538. The Poe cohort is below 300, so exact rates are directional. Free AI use is not observed.
Chart data
| Series | Share paying for at least three AI tools |
|---|---|
| Paid Poe subscribers | 46.5% |
| Other paid-AI subscribers | 5.2% |
How many Poe subscribers also paid for another AI tool?
Among paid Poe subscribers, 65.8% also paid at least one measured rival, versus 23.4% of other paid-AI subscribers. Poe subscribers averaged 2.57 paid AI tools compared with 1.30 for the comparison group. Claude appeared for 47.7% of Poe subscribers, Perplexity for 39.4%, and ChatGPT for 34.8%.
Did rival payments continue after subscribers adopted Poe?
Among 102 Poe subscribers who also paid a measured rival, 84.3% had another rival receipt at least 60 days after their first Poe payment, and 67.6% had paid a rival before Poe. A later receipt is evidence of continued observed payment, not confirmed ongoing usage, and the timing does not prove Poe caused subscribers to keep or drop another service.
Assumptions & Methodologies
We analyzed established U.S. e-receipt panelists from January 2024–August 2026. Paid Poe subscribers had a direct subscription, payment, or upgrade receipt; the cohort included 155 panelists. The comparison included 104,538 other paid-AI subscribers. AI tools were Poe, ChatGPT, Claude, Perplexity, Midjourney, ElevenLabs, and Mistral. The temporal check used 102 Poe subscribers with a paid rival and compared rival receipt dates with first Poe payment. Poe’s direct-billed cohort is small and may differ from app-store buyers. Free use is invisible, receipts measure payment rather than usage, and observed overlap or timing does not establish substitution or causality.