Similarweb: ChatGPT's worldwide desktop and mobile website visits fell 3.2% MoM to 1.43B in August, following ~10% drops from each of the previous two months
students are a large segment of usage Vishvanand Subramanian / @vishvanands : News outlets have moved on from Threads “unraveling” and now claim that ChatGPT is dying.. As always, citing SimilarWeb, the media says ChatGPT traffic has fallen for the 3rd consecutive month worldwide. But, looking at Google Trends (estimating search traffic), I see a large increase in August. … Dare Obasanjo / @carnage4life : It's hard to take any article that says web traffic to ChatGPT has been declining for 3 months seriously if it also doesn't point out they launched an iPhone app 3 months ago followed by an Android app about a month ago. Mastodon: Dare Obasanjo / @carnage4life@mas.to : It's hard to take any article that says web traffic to ChatGPT has been declining for 3 months seriously if it also doesn't point out they launched an iPhone app 3 months ago followed by an Android app about a month ago. — https://www.axios.com/... X: @wholemarsblog : why do you think that is? This entire hype cycle is built on the shoulders of ChatGPT so you guys better keep using it Carl Quintanilla / @carlquintanilla : “.. visits to the #ChatGPT website decreased by 3.2% .. in August, following approximately 10% drops from each of the previous two months.” @Reuters @knowledge_vital https://www.reuters.com/... Kenneth Li / @kenli729 : Exclusive: ChatGPT traffic slips again for third month in a row, recovery in sight https://www.reuters.com/... by @annatonger
Context & Ripple Effects
This extends an earlier warning sign: July tracking data showed declines in US ChatGPT visits and iOS downloads, making August a third consecutive month of lower web traffic rather than an isolated fluctuation. The earlier US usage slowdown also highlights that different measurement sources capture different surfaces of the product.
The reported figure is specifically for website visits, while the surrounding coverage notes the growing importance of app-based access. That distinction matters when interpreting whether lower web traffic reflects reduced use or a shift in where use occurs.
First-order effects
- Similarweb’s measure shows ChatGPT’s website audience contracting again in August, reducing the immediate web-traffic signal available to OpenAI and observers.
- The result puts greater weight on separating browser activity from app activity before treating website visits as a proxy for overall ChatGPT demand.
Second-order effects
- Usage trackers and media outlets will need to reconcile web, mobile-app, and search indicators; a single declining web series can produce an incomplete view of engagement.
- OpenAI’s distribution choices across browser and mobile become more consequential, because moving users between access points can change headline traffic metrics without a matching change in underlying usage.
Third-order effects
- Consumer AI products are likely to be evaluated through a fragmented set of engagement measures rather than one universal traffic number, complicating comparisons among services.
- If browser-to-app migration persists, the industry’s attention may shift from raw web visits toward cross-platform retention and monetization signals; this report alone cannot establish that migration.
The trend: Generative-AI consumer services are moving from novelty-driven web traffic toward a more fragmented, multi-surface measurement challenge.